Results of revision intramedullary nailing with and without auxillary plate in aseptic trochanteric and subtrochanteric nonunion

Only patients without clinical signs of local infection and without any evidence of infection in the anamnesis were recorded. Clinical signs of aseptic nonunion included an inability to perform a full weight load without pain or persistent instability in the fracture area. Radiologic signs of nonunion were defined as a lack of bony bridging in at least three of four cortices assessed on conventional radiographs in antero-posterior and lateral views. If conventional radiographs were inconclusive to diagnose nonunion, a computed-tomography (CT) scan was performed to determine the presence of nonunion. Preoperative biplanar radiographs and CT scans, and intraoperative, postoperative radiographs as well as the complete medical report were documented. The preoperative work-up for infection also consisted of blood samples including C-reactive protein and white blood cell values. Coding of the initial fracture was done independently by two senior orthopaedic surgeons. The AO/OTA classification (type 31 A1–3) was used for trochanteric fractures and the Seinsheimer classification (grade 1–5) for subtrochanteric fractures [16]. The currently accepted definition describes nonunion as a fracture that will not heal without further medical intervention regardless of the duration of treatment [1]. Nonunion was differentiated into atrophic and hypertrophic according to the radiographic pattern. Besides, identification of the previous implant and the potential need for any special equipment required for removal was conducted.

All surgical procedures were performed in a standardized manner under image intensifier control with patients in a lateral position on a radiolucent operating table according to the instructions provided by Friederichs et al. [17]. The surgical measures as well as the procedure and selection of implants were comparable in both treatment groups. The inserted fixation material was removed through the present approach. In the case of a varus or flexion deviation, the axis was corrected via the new implant. In each case, the medullary canal was reamed. If no new entry point was required, the same entry canal was overreamed slightly more posteromedially. The overreaming of the nonunion area was accomplished stepwise. A revision intramedullary nail as long and thick as possible was applied. The inserted revision nail was used in both treatment groups according to manufacturer's instructions (Gamma3 Nailing System, Stryker Corp., Kalamazoo, MI, USA; TRIGEN INTERTAN, Smith & Nephew Inc., Memphis, TN, USA). The revision implant in each case was selected by the treating surgeon. The aim was (1) to avoid the position of the previous intramedullary nail and (2) to improve stability by positioning the revision nail in stable bone depending on type and position of the previous nail. The nonunion area was debrided via limited exposure, followed by subsequent internal fixation using an angle stable 3.5 mm small fragment plate (LCP, Synthes GmbH, Oberdorf, Switzerland) [18]. The optional use of additional autologous cancellous bone harvested from the ipsilateral iliac crest was dependent on the discretion of the treating surgeon. It was based on the size and extent of the bone loss and the vitality and quality of the tissue in the nonunion area.

Intraoperatively, tissue samples were collected for histopathological and microbiological examination either from the reaming debris in group 1 or directly from the nonunion tissue in group 2 to exclude an inflammatory cause for nonunion development [19]. After sample collection, a single prophylactic antibiotic administration was applied based on the clinic protocol. Aftercare was conducted using the same rehabilitation protocol allowing early weight bearing in all patients [20]. Septic nonunion, nonunion based on pathologic or periprosthetic fractures, and nonunion treated by an external fixation procedure were excluded from this study. Besides, patients under 18 years of age and patients unable to provide written informed consent were excluded.

Patients were followed-up functionally and radiologically at 6 weeks, 3 months and 12 months after surgical revision in the outpatient department of the hospital. Functional outcome and patient satisfaction were observed using the Lower Extremity Functional Scale (LEFS) [21]. This score consists of 20 items, each with a maximum score of 4 points. The total possible score of 80 points indicates a high functional level of the lower extremity. Range of motion (ROM) of the hip joint was assessed using the neutral zero passage method [22]. Radiologic follow-up was assessed using anterior–posterior and lateral radiographs. Bone healing was defined as the ability to bear full weight without pain, stability at the former nonunion site, formation of bridging callus at all four cortices, and absence of fracture lines [23]. The cervicodiaphyseal angle (CDA) was measured between the femoral shaft axis and the femoral neck axis, normally ranging between 110° and 140°, with an average of 128° (female 127°, male 132°). Varus axis deviation is associated with an angle less than the normal range, and valgus axis deviation with an angle greater than the normal range.

Data were managed using Excel® for Windows® (Microsoft Corp., Redmond, WA, USA). IBM SPSS® Statistics for Windows (IBM Corp., Armonk, New York, USA) was used for statistical analysis. Test for normal distribution was performed using the Shapiro–Wilk test. Chi-squared test (gender, type of nonunion, nonunion healing, ROM, implant failure), Mann–Whitney test (Fracture type, radiographic classification, LEFS) and Student’s t test (age, time between initial and revision surgery) were used for statistical analysis. A result was considered to be statistically significant with p value < 0.05.