Outcomes of long retrograde intramedullary nailing for ...

8 downloads 56675 Views 425KB Size Report
Jai‑Gon Seo. Received: 7 June 2013 / Published online: 26 November 2013 .... placed and the prosthesis stable, a consensus on the best surgical treatment for ...
Arch Orthop Trauma Surg (2014) 134:47–52 DOI 10.1007/s00402-013-1890-7

TRAUMA SURGERY

Outcomes of long retrograde intramedullary nailing for periprosthetic supracondylar femoral fractures following total knee arthroplasty Sung‑San Lee · Seung‑Jae Lim · Young‑Wan Moon · Jai‑Gon Seo 

Received: 7 June 2013 / Published online: 26 November 2013 © Springer-Verlag Berlin Heidelberg 2013

Abstract  Background  The treatment of periprosthetic supracondylar femoral fractures following total knee arthroplasty (TKA) is challenging because of osteopenia and the limited bone available for distal fixation. The purpose of this study was to report the outcomes of periprosthetic supracondylar femoral fractures treated with long retrograde intramedullary nailing. Methods  We conducted a retrospective review of 25 patients who were treated with a long retrograde intramedullary nail for periprosthetic supracondylar femoral fractures following TKA. Clinical evaluation included range of motion of knee, Knee Society Score (KSS), Western Ontario and McMaster Universities Arthritis (WOMAC) score, and radiologic evaluation including time to union, coronal and sagittal alignment of femoral component, lower limb alignment, and implant loosening. The mean duration of follow-up after the fracture repair was 39 months (range 12–47). Results  All 25 fractures were united with a mean time of 12 weeks (range 8–20). At the last follow-up, the mean knee flexion was 111° (range 60°–130°), the mean KSS was 81.5 (range 50–100), and the mean WOMAC score was 30.2 (range 5–55). Four (16 %) of the 25 patients developed malalignment according to Rorabeck and Taylor criteria, but all patients had a knee flexion of more than 90°. Coronal and sagittal alignments of femoral component and lower limb alignment did not differ significantly between before and after the fracture repair. Complications included

S.-S. Lee · S.-J. Lim (*) · Y.-W. Moon · J.-G. Seo  Department of Orthopedic Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon‑Dong, Gangnam‑Ku, Seoul 135‑710, South Korea e-mail: [email protected]

the loosening or breakage of distal interlocking screws in three patients. No deep infection or prosthesis loosening was detected at the last follow-up. Conclusions  Surgical treatment of periprosthetic supracondylar femoral fractures following TKA with long retrograde intramedullary nailing resulted in high union rates and encouraging functional outcomes. Keywords  Total knee arthroplasty · Supracondylar femoral fracture · Long retrograde intramedullary nailing

Introduction An increase in the number of patients with total knee arthroplasty (TKA) has occurred because of the increasing life expectancy and functional demands of the elderly population [1]. Due to higher activity and incidence of osteoporosis, the number of periprosthetic fractures is also very likely to increase [2, 3]. The prevalence of periprosthetic supracondylar femoral fractures following TKA ranged from 0.2 to 2.5 % [4, 5]. The treatment of periprosthetic supracondylar femoral fractures following TKA is challenging because of the frequent occurrence of osteopenia, limited bone available for distal fixation, and potential for complication related to arthroplasty, such as femur implant notching [6]. The conventional open plating technique causes vascular disruption which increases the risk of malunion and mechanical failure [7]. Currently, two types of fixation techniques including retrograde intramedullary nailing and locked plating are being used to treat periprosthetic supracondylar femoral fractures [7–11]. Both techniques have demonstrated significant fracture union rate and functional outcomes, while each technique has relative benefits and potential

13

48

pitfalls [8]. The advantages of retrograde intramedullary nailing are that it is a load-sharing device compared with a locked plate, providing a greater stability and smaller approach window needed to perform the procedure [8]. Potential disadvantages of retrograde intramedullary nailing include knee sepsis, stiffness, and hyperextension of the femoral component due to an incorrect nail entry point [9]. The advantages of locked plating are that it is compatible with the pre-existing knee prosthesis and provides improved pull-out strength even in an osteoporotic bone. The potential disadvantages of locked plating techniques are the longer operative times and irritation of the iliotibial band and soft tissues of the knee, necessitating their removal [7, 10]. While some authors compared both techniques using clinical and radiologic results, whether one technique was better than the other has not been demonstrated [10, 11]. The purpose of this study was to report the clinical and radiologic results of patients who had undergone long retrograde intramedullary nailing as the treatment of supracondylar periprosthetic femoral fractures following TKA.

Material and methods With approval from the institutional review board, we conducted a retrospective review of all patients who had sustained supracondylar femoral fractures following TKA at our institution from February 2005 to January 2011. Inclusion criteria were supracondylar femoral fractures following TKA with retrograde intramedullary nailing fixation. Exclusion criteria were: antegrade intramedullary fixation or utilization of locking or non-locking plates and follow-up of less than 12 months, pathologic fractures, or periprosthetic femoral fractures with definite loosening of prosthesis. Thirty-two patients underwent surgical treatment for 32 supracondylar femoral fractures following TKA during the study period. A total of seven patients were excluded because of internal fixation other than retrograde intramedullary nail (five patients) and loss to follow-up (two patients). Thus, the final study group consisted of 25 patients (25 knees). The patients included 24 women and 1 man, with a mean age of 71 years (range 62–86 years) at the time of fracture repair. The average time interval from TKA to fracture repair was 52.5 months (range 43 days–14 years). The mean duration of follow-up was 39 months (range 12–47). The original diagnoses were osteoarthritis in 24 patients (96 %) and rheumatoid arthritis in 1 patient (4 %). Radiographs consisting of anteroposterior and lateral views of the initially injured distal femur were obtained in each patient. Additional computed tomography scans with coronal and sagittal reconstructions were performed at the surgeon’s discretion for

13

Arch Orthop Trauma Surg (2014) 134:47–52 Table 1  Classification according to AO/OTA [12] Classification

Description

Number of knees

33 A1 33 A2

Simple Metaphyseal wedge

20 (80 %) 4 (16 %)

33 A3

Metaphyseal complex

1 (4 %)

assessing fracture pattern and implant loosening. According to the AO/OTA (Arbeitsgemeinschaft Osteosynthese/ Orthopaedic Trauma Association) system [12], all fractures were classified as 33 Type A fractures (Table 1). Fractures were also classified according to Rorabeck and Taylor classification [13] (Table 2) as well as Su et al. [14] classification (Table 3). In all cases, patients were operated in the supine position on a radiolucent table. A longitudinal anterior knee incision was made using the prior arthroplasty scar. Medial parapatellar arthrotomy was also used, similar to that in the previous arthroplasty. The intercondylar notch of the femoral component was exposed and a guidewire was inserted into the opening of the distal femur. Care should be taken not to place an overly posterior nail entry point in the intercondylar notch which might produce hyperextension deformity of the femoral component. The fracture was reduced by manual longitudinal traction and a sterile bolster was placed under the distal thigh to control sagittal plane alignment. When the fracture could not be reduced by manual traction alone, several instruments were used to assist in fracture reduction including reduction forceps, ball spike pushers, or Steinman pins. When necessary, a percutaneous cerclage wiring technique was used to achieve and maintain fracture reduction before nail insertion. A standard cannulated reaming technique was carried out under fluoroscopic guidance. Reaming was performed until cortical chatter was encountered and a nail that was 1.0–1.5 mm less in diameter than the final reamer was advanced into the medullary canal following a guidewire. A retrograde femoral nail (Solco Biomedical, Seoul, Korea) was used in all of them. For each procedure, we inserted a full-length, long intramedullary nail to the level of the lesser trochanter to prevent so-called ‘windshieldwipering’ by crossing the isthmus and thereby improving stability. Using an aiming arm with an attached nail insertion guide, two or three distal interlocking screws were inserted as was determined in the previous biomechanical study to provide better axial and torsional stability [15]. One or two proximal interlocking screws were inserted using a freehand technique. A closed-suction drain was routinely used. Active-assisted range-of-motion exercises were started on postoperative day 2 after removal of the drain. All patients were allowed to stand on the second postoperative day and to progress to partial weight-bearing

49

Arch Orthop Trauma Surg (2014) 134:47–52 Table 2  Classification according to Rorabeck and Taylor [13]

Classification

Description

Number of knees

Type 1 Type 2

Undisplaced fracture; prosthesis intact Displaced fracture; prosthesis intact

1 (4 %) 24 (96 %)

Type 3

Displaced or undisplaced fracture; prosthesis loose or failing

0 (0 %)

Table 3  Classification according to Su et al. [14] Classification Description

Number of knees

Type 1 Type 2

Fracture proximal to the femoral component Fracture originating at the proximal end of the femoral component and extending proximally

0 (0 %) 19 (76 %)

Type 3

Fracture in which any part of the fracture line can be seen distal to the upper edge of the anterior flange of the femoral component

with crutches as tolerated. The patients were allowed full weight-bearing after 6–8 weeks. Clinical and radiographic evaluations were performed preoperatively, and at 6 weeks, 3, 6 months, and 1 year postoperatively, and then annually. Clinical evaluations were performed using ranges of motion of the knee, Knee Society Score (KSS) [16], and Western Ontario and McMaster Universities Arthritis (WOMAC) score [17]. Radiographic evaluations were performed using standardized anteroposterior and lateral radiographs of affected knees taken postoperatively, during hospitalization, and at each follow-up visit. All radiographs were digitized using PathSpeed software (General Electric Inc, Milwaukee, WI) and reviewed by a single independent observer, who did not participate in clinical care. Radiographic parameters included time to union, coronal and sagittal alignment of femoral component, and implant loosening. Union of fracture was defined as the formation of bridging callus across the fracture site on each of the anterior–posterior and lateral radiographic views, and nonunion was defined as no evidence of sufficient callus 6 months after the fracture. A malalignment was defined as varus/valgus of the distal fragment >5°, flexion/extension >10°, or shortening >2 cm [13]. Implant loosening was defined by continuous or progressive radiolucent lines or by migration of any component. To evaluate coronal and sagittal alignments of femoral components and the lower limb alignment, we used α-angle, γ-angle, and femorotibial angle (Fig. 1) [18]. We also compared the femoral component and lower limb alignments before periprosthetic fractures with those after the index procedure. Statistical analysis was performed using the SPSS statistical software system (version 15.0; SPSS, Chicago, IL). The paired t test was used to compare femoral component alignments immediately before periprosthetic fractures with those at the final follow-up. P values of