1 Introduction

Patella fractures represent a relatively small proportion of orthopaedic injuries seen in skeletally mature patients. With estimates around 1%, these fractures most commonly follow direct trauma, as is the case with falls or dashboard injuries.^1,2 They can also occur via an indirect mechanism, due to the tensile forces applied to the patella via the extensor mechanism of the lower extremity.³ Fracture patterns associated with the patella typically depend on the mechanism of injury, associated force, and underlying bone quality.⁴ Regardless of the degree of comminution, a majority of patella fractures have a transverse component as a major feature. Those with a disrupted extensor mechanism, and/or those with articular incongruity generally require open reduction and internal fixation. Multiple operative techniques are described. Challenges of fixation include a limited anterior soft tissue envelope, minimal surrounding periosteum, and most importantly the forces of tension that make successful internal fixation difficult, regardless of the construct employed. Complications, including early failure of fixation, are relatively common, compared to other fractures.

For transverse fractures of the patella (Fig. 1), modified tension band constructs have been considered the gold standard for operative fixation, based on biomechanical principles of fixation.^5–8 While variations exist, they function, in part, by converting the distraction forces applied to the patella via the quadriceps and patella tendons into compressive forces at the articular surface.⁹ Several techniques have been described, varying in how they achieve fixation in the bone and how the tension band is applied. The traditional model employs two longitudinal K-wires acting as anchors for a tension wire passing transversely through the fibers of the quadriceps and patella tendons, across the dorsal cortex of the patella. This has been shown to achieve relatively uniform compression across the fracture. By passing transversely through the quadriceps and patella tendons, the wire is able to take advantage of the entire surface area of insertion, held fast by Sharpey's fibers (Figs. 2 and 3). Because the wire to tendon/bone interface is so strong, the weak link is the K-wires, which are essential in holding the wires anchored in place. The mode of failure of this tension band construct is migration of the k-wires and subsequent loss of anchorage of the tension band wire (Fig. 4).

Close

Fig. 1 AP and lateral radiographs of simple transverse Patella fracture.

Export to PPT

Close

Fig. 2 Rendition of the Kirschner wire Tension Band Construct.

Export to PPT

Close

Fig. 3 AP and Lateral Radiographs of the Kirschner wire Tension Band Construct.

Export to PPT

Close

Fig. 4 AP and Lateral Radiographs of the Kirschner wire Tension Band Construct Failure.

Export to PPT

In part to address this mode of failure, constructs have been introduced that utilize two longitudinal, cannulated, partially threaded screws with a wire or suture passed through the central cannulation, in place of the K-wires (Figs. 5 and 6). This model was intended to improved fixation, based on the thread-bone interface that screws provide. In addition, it was felt to be less prominent and therefore less likely to lead to symptomatic hardware and subsequent hardware removal (Fig. 7).^5,10,11

Close

Fig. 5 Rendition of the cannulated screw tension band construct.

Export to PPT

Close

Fig. 6 AP and lateral radiographs of the cannulated screw tension band construct.

Export to PPT

Close

Fig. 7 AP and lateral radiographs of the cannulated screw tension band construct failure.

Export to PPT

However, the switch from K-wires to cannulated screws did more than simply change the anchor points for the wire. It also changed the way the wire functions. When passed around longitudinal K-wires, the tension band wire gets its fixation on the bone from the broad insertions of the tendons into bone, through Sharpey's fibers. However, when the wire is passed through parallel screws, the orientation of the wire is changed 90°. The wires, running longitudinally, come into direct contact with the bone at the poles, and apply the force of fixation through very limited points of contact. This difference theoretically alters the potential modes of failure of the two methods. The K-wire and tension band wire constructs most often fail with primary loss of K-wire position, and subsequent loss of tension wire function.^4,7,8 Conversely, those with cannulated screws would more likely fail as a result of bone fragmentation at the areas of concentrated force, where the wires contact the bone, as the screw threads are less likely to lose positional fixation compared to K-wires.¹ This proposed difference led to the development of a hybrid fixation construct, that builds on the strength of each technique.

These methods are widely employed, and yet studies continue to report failure rates around 10%.^{5–7,12–16} This record of performance, as well as our own observed failures, was the impetus for an evolution in fixation strategy.

The purpose of this study is to evaluate the effectiveness of a fixation construct that evolved from both above-described techniques to avoid fixation failure, recognizing the strengths and weaknesses of each. We have therefore established construct failure as the primary clinical outcome measure. This simple technique involves two parallel, fully threaded cortical screws, placed longitudinally in the patella, with enough proximal and distal overhang to allow them to act not only as screws, but as posts, around which the tension band wire is anchored. The screws can gain independent fixation in the proximal and distal poles, but more importantly, because they are threaded, they do not have a tendency to migrate or become displaced. Because the tension band wire is passed transversely through the quadriceps and patellar tendons, the forces applied through the wire are shared along a broad insertion, and anchor to the bone via Sharpey's fibers, which is among some of the strongest tissues in the body (Figs. 8 and 9).

Close

Fig. 8 Rendition of the cortical screw tension band construct.

Export to PPT

Close

Fig. 9 AP and lateral radiographs of cortical screw tension band construct.

Export to PPT

2 Methods

Following Institutional Review Board approval, an electronic medical record database query was performed to identify all cases of patella fractures treated with open reduction internal fixation using the previously described technique by a single surgeon at a single institution between 2013 and 2019. Excluded were cases of suture only fixation through bone tunnels, revision surgery, or those who did not have a minimum of 6 month follow up. Operative reports were used to collect intraoperative and demographic variables that included case duration, fixation technique, American Society of Anesthesiologists physical status score (ASA), and patient age and sex. Postoperative imaging was also reviewed to confirm fixation techniques. Body mass index (BMI), Charlson Comorbidity Index (CCI) scores, patient smoking status, and postoperative deep vein thrombosis (DVT) prophylaxis regimens were also determined through electronic medical record review. The postoperative course was reviewed, and any complications, revisions, or subsequent surgeries were documented.

Data was collected and analyzed using the Statistical Package for the Social Sciences (SPSS Inc, Ver 26.0). Demographic and outcome characteristics of the study population were reported using descriptive statistics. Outcome measures were established to assess the relative risks associated with BMI, CCI score, smoking status, and patient age and sex. The reported outcomes included surgery duration, construct failure (defined as breakage, migration, or cut out in the setting of a persistent radiographic nonunion and corresponding clinical symptoms attributable to said nonunion or hardware failure), symptomatic removal of hardware (ROH), and revision surgery. Relative risks (RR) for construct failure were calculated and potential predictors of construct failure were identified using binary logistic regression analysis. Patients lacking clinical follow-up were excluded from statistical analysis. Statistical significance for all testing was established at p < 0.05.

3 Results

Forty-nine patients were identified and included in the final study cohort. An additional eleven patients were identified but lacked adequate postoperative follow-up. Median follow-up was 52.7 weeks (Mean: 75.4 weeks; SD: 54.9; range: 27–267.7 weeks). There were 15 males and 34 females in the cohort with a mean age of 61.7 years old (SD: 15.4; range: 21–84 years old). The mean CCI score was 0.47 (SD: 0.74; range: 0.0–2.0) and the mean age-adjusted CCI score was 3.2 (SD: 1.7; range: 0.0–7.0). Mean BMI was 26.7 (SD: 6.3; range: 18.6–48.9). Thirteen patients (26.5%) had a positive smoking history, with four (8.2%) current smokers and nine (18.4%) former smokers. Mean ASA was 2.1 (SD: 0.61; range: 1.0–3.0). Mean surgery duration was 91.3 min (SD: 44.1 min; range: 37–274 min. Enoxaparin was the most frequently utilized DVT prophylactic, used in 16 patients. Aspirin was used in 14 patients. Other documented DVT prophylaxis regimens included, aspirin and dipyridamole (n = 1), and aspirin and clopidogrel (n = 1). The rate of construct failure was 3/49 (6.1%). This included two mechanical hardware failures, one the result of a fall, and one occurring in a patient who was noncompliant with our postoperative extremity protocol that involves wearing a knee brace locked in extension for a minimum of 8 weeks. The third construct failure was secondary to deep infection. All three failures required revision ORIF. There was also one medical complication in the cohort, which was an episode of postoperative cardiac arrest secondary to lidocaine administration (Table 1). Female sex, BMI ≥30.0, age ≥65 years old, ASA ≥3, age-adjusted CCI ≥4, and positive smoking history carried no significantly increased risk for construct failure (Table 2) and were not significant predictors for construct failure (all p > 0.05). Reoperations also included 30 patients (61.2%) who underwent planned removal of symptomatic hardware during the study period.

4 Discussion

The patella is the largest sesamoid bone in the body and serves to optimize the quadriceps function in the lower extremity.⁴ It augments the extensor and counter-flexion mechanisms of the knee by providing a longer lever arm for the quadriceps muscle group.^17–22 Accordingly, the patella can be subjected to compressive and tensile forces beyond 3000 N, which far exceed the standard body weight of even a significantly overweight general population.²² These unique biomechanical forces highlight some of the challenges of fixing patella fractures, but also underscore the necessity to treat them definitively with surgery.

Despite the fact that outcomes following internal fixation of patella fractures have improved significantly over time, failure rates still remain quite high.^14–16 Lin et al. demonstrated a failure rate of 7.7% with a cannulated-screw modified tension-band construct, while Bonnaig et al. described a nonunion rate of 10% in the K-wire modified tension-band construct.^13,15 Though the failure rate of our study (6.1%) did not differ significantly from what is seen in the literature, our failure rate included all patients requiring revision fixation for osseous nonunion and hardware failure regardless of how it developed. This included one patient who refractured after a fall and another who did not wear her extension brace postoperatively. If those two patients were excluded, the failure rate for the cortical screw tension band construct falls to 2.0% (1/49). However, similar instances of noncompliance could likely be cited among some of the reported failure rates in the literature.

The patella is relatively lacking in its soft tissue envelope, leading to a somewhat unforgiving environment for surgical implants. It has a relatively sparse periosteal layer, and thus healing can be challenging. While the rate of infection in our study was only 2.0%, reports from other fixation constructs approach 10%, contributing to the significant morbidity associated with these injuries.^{14–16,23–25}

5 Conclusions

A modified tension band construct that utilizes cortical screws is a viable option for treating transverse patella fractures. Although this construct is associated with a high rate of hardware removal, the overall failure rate is below historical rates for other modified tension band constructs, especially after considering recurrent trauma and patient noncompliance as contributing factors. Further investigation is necessary to validate this technique.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.