“Revision of subtrochanteric femoral nonunions after intramedullary nailing with dynamic condylar screw”

Subtrochanteric femoral fractures account for approximately 25% of all hip fractures and have a bimodal age and sex distribution [1]. The subtrochanteric region is defined as the area between the lesser trochanter and the femoral isthmus, which is five centimeters below the trochanter [2, 3]. The femur is exposed to biomechanical forces due to the osseous anatomic conditions in the region, as well as the skeletal muscles surrounding the hip joint. Shear forces specifically affect the cancellous bone of the proximal femur, whereas bending forces especially affect the cortical bones of the subtrochanteric area and the shaft [4–6]. Nonoperative treatment of subtrochanteric femur fractures is associated with a high rate of complications [7]. Although the most recent literature does not emphasise the superiority of intramedullary fixation with nail systems (IMN), it should be however considered as a primary treatment option more particular in elderly patients [8, 9]. Nevertheless, extramedullary fixation with a 95° angled blade plate and plate-sliding screw systems also has significant value in current surgical therapy [1]. IMN seems to have biomechanical advantages due to a shorter lever arm compared to the extramedullary components. However, these advantages are partially outweighed by a difficult closed reduction [10]; screw cut outs, screw migration peri-implantary femoral fractures, malunion and nonunion are not uncommon [11–13]. General risk factors [14], as well as local risk factors due to the magnitude of trauma, can lead to delayed bone healing or nonunion in as many as 20% of subtrochanteric femur fractures [15–18]. A significant increase in nonunions of the proximal femur is expected in the future, especially in patients older than 80 years. This is due to demographic changes and the consequently greater number of proximal fractures of the femur in the population [19–21]. Nonunions of the subtrochanteric region of the femur after prior IMN can be difficult to address. Implant failure and bone defects around the implant significantly complicate the therapy, and complex surgical procedures with implant removal, extensive debridement of the nonunion site, bone grafting and reosteosynthesis usually become necessary [22]. The Dynamic Condylar Screw (DCS; Synthes, Bettlach, Switzerland) has been designed for the internal fixation of fractures of the distal and subtrochanteric regions of the femur and has superior biomechanical properties compared to the blade plate [23–25]. The purpose of this study was to evaluate a series of DCS treatments of patients with subtrochanteric femoral nonunions following IMN regarding healing rate, secondary surgeries and implant-related complications.

Between January 2002 and July 2017, 40 patients fulfilled the inclusion criteria and were included in the study. The average length of the follow-up period was 26.3 months (range 6–173 months), with a median of twelve months. The mean age of the 24 women and 16 men upon admittance to our clinic was 65.4 years (range 34–91 years). Regarding clinical constellations, initial fractures were treated with a gamma nail (fa. Stryker) in 17 patients, and a proximal femoral nail (PFN; fa. Zimmer) was used in eight patients. Fifteen fractures were treated by Targon-, Sirius-, Trigen-, Vero-, Friedel or Orthofix intramedullary nails. Among these cases, ten of the 40 patients (23.3%) had already undergone revision surgery after nailing, including for dynamization of the nail, bone grafting, additional augmentation plate insertion or exchange nailing (Fig. 2). A total of three patients had undergone more than one revision surgery. Of the 40 total nonunions, nine (22.5%) were associated with implant failure at admission to our hospital (Fig. 5). The 40 aseptic nonunions were classified as atrophic in 35 (87.5%) cases and as hypertrophic in five (12.5%) patients.

The index surgery was performed at an average of 9.88 months (± 6.7 months) and a median of nine months after the initial nailing. The mean operation time was 154.25 min (range 70–265 min). In 18 patients (45%), an additional LCDCP was used. With the exception of ten patients, an autologous cancellous bone graft taken from the anterior iliac crest was used. In five patients, a demineralized bone matrix was applied instead of an autologous bone graft. The remaining five patients without additional bone grafting had hypertrophic nonunions. The preoperative CCD of the 40 patients prior to the index procedure was 118° (range 101° – 131°). The mean postoperative CCD was calculated to be 128° (range 114° – 142°) with a mean valgus correction of ten degrees. The mean length of the hospital stay was 15.08 days (range 6–40 days). On average, three units of red blood cells were transfused during the hospital stays (range 0–12 units).

Revision surgery was performed at a mean of 4.9 months after the index procedure (range 4 days – 12 months). In cases of a persistent nonunion with hardware failure (n = 9), removal of the plate and a revision procedure using the same implant again or a double plate construct (DCS + anterior LCDCP; n = 4) and iliac bone crafting was performed after a mean of 5.9 months (range 1–15 months). In one case, a patient developed a femoral head necrosis with screw cut-out and was treated with THR. In the case of a persistent nonunion without implant failure (n = 1), further revision with additional anterior plating using LCDCP and iliac bone crafting was performed. Due to infections (n = 2), local surgical debridement was performed leaving the plate in situ with subsequent antibiotic therapy. Implant removal was performed in seven patients after a mean follow-up period of 14 months (range 6–24 months) due to a planned foreign body removal or local pain sensation. Of the 13 patients, a subcategory of eight patients (20%) needed a second additional surgery. Second revisions were performed a mean of two months (range 6 days – 8 months) after the first revision. Second revision surgeries were necessary due to implant failure (n = 4), persisting nonunion without implant failure (n = 1) and a deep infection requiring surgical debridement (n = 3). In the case of a hardware failure, the same revision procedure with double plating was repeated. In the three patients with infections, debridement was repeated in two patients. In one case, resection arthroplasty resulting in a Girdlestone situation was performed. The nonunion without implant failure was treated with an iliac bone graft and demineralized bone matrix. A total of three of the 13 patients (7.5%) patients needed one or more additional surgeries, leading to two cases of osseous consolidation and one case of persisting nonunion.

At the final follow-up visit, a full range of motion of the hip was seen in ten of the 40 (25%) hips. The other 30 hips (75%) had a limited range of motion, with a minimum of 90° of hip flexion. A total of 15 patients (37.5%) could walk without any walking aids, nine patients (22.5%) walked with forearm crutches and eight patients (20%) walked with a walking frame. Two patients used a wheelchair (5%). One patient was bedridden (2.5%). In five patients, postoperative mobility could not be evaluated due to a lack of data.

The purpose of this study was to evaluate a series of DCS treatments of subtrochanteric femoral nonunions following IMN regarding the healing rate and implant-related complications. To the best of our knowledge, our patient series of 40 subtrochanteric femoral nonunions represents the largest study reporting on a standardized salvage DCS procedure after failed intramedullary nailing.

Nonunions of the subtrochanteric region of the femur are uncommon; however, they are difficult to treat. In addition to general risk factors [14], mechanobiological properties such as the stability of fracture fixation influence the cellular processes in the healing tissue [30]. Cephalomedullary interlocking nails have improved the results of subtrochanteric fractures with high union rates and have decreased the incidence of fixation failure [1, 31–33]. However, use of IMN remains technically demanding and requires proper reduction and correct positioning of the implant [34]. Malreduction in either the coronal (comminution of the medial cortex) or sagittal plane leads to prolonged time to union for subtrochanteric fractures [10, 35, 36]. Percutaneous reduction maneuvers and minimally invasive techniques can be useful in obtaining proper reduction [3]. Immediate unrestricted weight bearing appears to be a safe postoperative regime for subtrochanteric femoral fractures treated by IMN [1]. Early mobilization can not only lower mortality and morbidity rates but also impel functional recovery in patients with proximal femur fracture [37, 38]. If nonunion occurs after intramedullary nailing, options for reconstruction include dynamization of the nail, exchange nailing, bone grafting, augmentative plating, nail removal and plating or prosthetic replacement. The length of the proximal fragment, femoral deformities and defects in the femoral bone stock guide decision-making. The presence of hardware and poor bone stock from prior fixation attempts can compromise stable fixation [39]. In the treatment of subtrochanteric femoral nonunion, a variety of implants have been used with variable success. The treatment of subtrochanteric aseptic nonunions is rarely reported, and sample sizes in published studies have been small [39]. Barquet et al. [40] presented a study of 29 patients with subtrochanteric nonunions managed by the removal of previous implants (n = 9 after nailing; n = 20 after open reduction and internal fixation (ORIF) with blade plate, dynamic compression plate or DCS), open correction osteotomy if needed (n = 3), secondary osteosynthesis with a gamma nail (overreaming to provide biologic augmentation) and bone grafting in case of loss of bone substance (n = 5) [40]. Of these nonunions, 23 (88%) healed after the index procedure. Except for one case (96%), all nonunions healed in a mean of seven months. In total, three implant failures (11.5%) were detected. Severe comminution and fragment diastasis or varus alignment were mentioned as influencing factors for persisting nonunion [41]. Wu treated 21 patients with subtrochanteric nonunions after ORIF or IMN (all described patients remained unhealed after one to six previous surgical treatments) with locked nail stabilization [42]. The patient age ranged from 19 to 56, with a median of 36 years. No patients required additional surgery with a healing rate of 100% after one year. Nail dynamization alone offers a minimally invasive treatment option for patients without unacceptable bony deformity or limb shortening. Dynamization has been shown to be effective [43, 44], but its effectiveness according to subtrochanteric nonunions has not been proven yet. In a study including 19 patients, Kang et al. [22] showed that the union rate with the exchange of previous implants, in addition to the complete removal of fibrous tissue and bone grafting, was better than in those with retained hardware in the treatment of subtrochanteric nonunion (ten vs. nine) [22]. In contrast to the abovementioned techniques that rely on renailing or nail dynamization, plate revision of the subtrochanteric femoral nonunions has also been reported. Focusing on the enhancement of mechanical and biological preconditions (subsumed as the diamond concept [45]), Giannoudis et al. [46] presented a study of 14 subtrochanteric nonunions after initial IMN revised with ORIF (95 degree angle blade, n = 11) or IMN (Affixus® Hip Fracture nail, n = 3) [46]. In total, one revision surgery due to a blade plate failure was necessary. Finally, all 14 nonunions healed after an average of 6.8 months. De Vries used blade plating in 33 subtrochanteric nonunions (mean age of 53 years) and reported a healing rate of 96.9% (n = 32) [47]. The average time to healing was five months. According to the functional outcome of the united nonunions measured by the Merle d’Aubigne score [48], ten patients scored as excellent, 15 scored as good and seven as fair outcomes. In total, nine postoperative complications (27.3%) after the index surgery were mentioned. These complications required six revisions, including three “minor” revisions such as one postoperative hematoma that required drainage, one protruding tip of the plate for partial removal of the implant and one superficial wound infection requiring debridement. In total, three “major” complications led to revision surgery including one collapse of the femoral head, one implant failure leading to THR and one refracture after implant removal leading to a revision with IMN and ultimately leading to union.

According to the current literature, there is no strong evidence to support the use of either IMN or extramedullary devices in the revision of subtrochanteric nonunions. Despite the development of new implants and increasing knowledge of nonunion, the treatment of postoperative complications and persisting nonunion still presents a challenge. To date, no prospective randomized study has been published; therefore, level IV studies are currently the best available evidence [49]. We presented a large series of late plate fixations of subtrochanteric femoral nonunions following IMN. In our current study, the final healing rate was 92.5% (37/40). A significant number of the initial nonunions (87.5%) were classified as atrophic nonunions. Therefore, fracture healing seemed to be disturbed more by biology than by biomechanics. Internal fixation causes significant additional damage to soft tissues in settings where bone healing is already compromised by prior procedures and implants [50]. Positive results have been reported for closed reduction and biologic plating by DCS for treating subtrochanteric fractures [51, 52]. In contrast, extensive debridement of the atrophic nonunion sites with bone grafting [53, 54] is one of the key factors in treatment of a nonunion. Therefore, complete exposure and opening of the nonunion site is necessary for successful surgical nonunion treatment. Additionally, in cases where significant deformities require proper anatomic reduction (for patients with very short proximal fragments with complex deformities or large bone defects), the theoretical advantages of biological plating or exchange nailing are refuted.

Our data series has a relatively high number of postoperative complications leading to revision surgery (13/40). Rosso et al. [55] and Kulkarni et al. [52] reported a high implant failure rate (26%) in elderly patients (> 50 years) using DCS for unstable subtrochanteric and intertrochanteric femoral fracture treatment. Most of the abovementioned studies included younger patients [42, 47]. The mean age in Giannoudis’s study was 68.4 years. Except for one 63-year-old patient, all of the mentioned complications (n = 6) occurred in patients older than 74 years [46]. Our study includes a large number of elderly patients. The average age of patients upon admittance to our clinic was 65.6 years. In total, 85% patients were 50 years old or older. Most of the patients were older than 60 years (64.6%) or 70 years (41.7%) at admittance. Thus, age was not a significant factor for revision surgery in our study because both groups (Groups 1 and 2) seemed to belong to a high-risk group suffering from implant failure or complications. For these patients, less favorable results and high implant failure rates have been reported previously [56]. Cement-augmented techniques have been described and might be possible alternatives in fracture and implant fixation; however, no study has explored subtrochanteric femoral nonunion treatment [57, 58]. Hip arthroplasty after failed fixation of trochanteric and subtrochanteric fractures might be another alternative; however, several studies highlighted the challenge of the procedure with comparable or even higher complication rates [59, 60]. The strengths of the present study are the relatively large population compared to other studies focusing on nonunions after IMN and the use of one type of implant as the revision procedure.

We acknowledge the limitations of the study. First, only retrospective data gathered from a local electronic database without a defined follow-up protocol is presented. Second, there is a heterogeneity of the study cohort. We included all participants with subtrochanteric nonunions following IMN regardless of the type of initial trauma, type of nonunion and the number of prior revisions. Although there was no significant impact on bony healing in any of these categories, in our cohort group that might be a result of subgroup analysis with small group samples. Third, the index procedure was performed over the course of 15 years observational time by 18 different surgeons with varying levels of experience. Fourth, the 40 included patients had been referred to our hospital for nonunion treatment. Additional information about inital fracture treatment was limited and the rate of nonunion after fracture treatment was not available.

This study shows that treatment of subtrochanteric nonunions is still a challenge. Despite high overall healing rates, a significant number of patients needed secondary surgical revisions, especially among elderly patients.