Treatment of aseptic nonunion after intramedullary nailing fixation with locking plate

Abstract

Objective:  To evaluate the safety and efficacy of locking plate combined with bone grafting in the treatment of aseptic nonunion following intramedullary nailing fixation of fractures of the long bones.

Methods:  Thirty‐eight consecutive patients treated in our hospital between January 2004 and December 2006 were included in this retrospective study. The nonunions included 20 femurs, 15 tibias, and 3 humeri. The duration of non‐union ranged from 6 to 84 months and 21 (55.3%) of them were located around the metaphysis of the affected long bones. There were 12 women and 26 men with a mean age of 39.2 years (range, 9–70 years). Locking plate combined with bone grafting was the procedure chosen to treat every case of nonunion in this series. The clinical outcomes were evaluated.

Results:  All patients were followed up for 6–20 months (average 11.6 months). After locking plate fixation combined with bone grafting, union was achieved in all cases, the average healing time being 5.3 months (range, 4–8 months). Infection of the superficial incision occurred in three cases (7.9%) and delayed healing of the incision in one case, all of which healed with no further complications. The function of the adjacent joints was excellent to good in 30 patients (78.9%), fair in 7 (18.4%) and poor in 1 (2.6%) after follow‐up.

Conclusion:  Locking plate fixation combined with bone grafting is a highly effective treatment for aseptic nonunions of the long bones after intramedullary nailing fixation, especially in the case of metaphyseal nonunion.

Introduction

Intramedullary nails have been widely used to treat long bone fractures, most of which heal within the expected time. However, in some cases union fails to occur by the expected time due to infection, screw and nail breakage and deformation ¹ , ² , ³ . After treatment with intramedullary nails, fractures located around the metaphyses of long bones are especially prone to nonunion. Aseptic nonunion is relatively common, and one of the most severe complications in the management of tibial and femoral fractures ⁴ . The treatment of nonunion is one of the greatest challenges in fracture care ⁵ . Some investigators have reported that treating nonunion, mainly in cases of tibial and femoral shaft nonunions, by changing the intramedullary nails or plating substitution, results in a rate of union ranging from 40% to 100% ⁴ , ⁶ , ⁷ , ⁸ , ⁹ . However, changing the intramedullary nail is not a reliable fixation technique for those nonunions near the metaphysis. Recently, Hailer and Hoffmann have reported one case of nonunion treated with less invasive stabilization system (LISS), in which anatomical pre‐shapes and locked screws with angular stability are used both in the shaft and the condylar region ¹⁰ . This prevents the screw from backing out and provides better fixation for the ununited bone. What about the efficacy of this procedure for nonunions after treatment by intramedullary nailing? Reports in the literature are rare. The purpose of this study was to evaluate the safety and efficacy of locking plate combined with bone grafting in the treatment of aseptic nonunions following treatment of fractures of long bones by intramedullary nailing fixation.

Materials and methods

Patient data

Thirty‐eight consecutive patients treated in our hospital between January 2004 and December 2006 for femoral, tibial, or humeral nonunions following intramedullary nailing fixation were included in this retrospective study. There were 12 female and 26 male patients with a mean age of 39.2 years (range, 9–70 years), and the nonunions were in 20 femurs (10 in the shaft, 3 in the proximal and 7 in the distal part), 15 tibias (4 in the shaft, 5 in the proximal and 6 in the distal part) and 3 humeri in this series. Twenty‐one cases (55.3%) were located around the metaphysis of the affected long bone.

The details of the original fractures and types of nail used were not known. Ten cases (26.3%) had been subject to more than two surgeries and 28 (73.7%) had undergone only one previous surgery. According to the Judet classification of nonunion ¹¹ , there were 9 hypertrophic nonunions, 10 avascular nonunions and 19 atrophic nonunions. The possible causes of nonunion for these cases were breakage of screw nails in 2 (5.3%), loosening of screw nails in 19 (50%) and breakage of intramedullary nails in 17 cases (44.7%). Pseudarthroses had formed in 15 cases. There were not malformed fractures in all cases. The mean time from the first surgery was 16.2 months (range, 6–84 months).

Surgical procedure

Under anesthesia, the old incision was reopened and the incision carried down sequentially to expose the site of nonunion. After removing any swelling osteotylus from the fracture surface and inserting the locking plate, general screws were used to fix the plate in temporarily and the locking plate compressed with bone. Locking screws were used to fix the locking plate when line of force and position were satisfactory as shown by X‐ray.

Two non‐locking screws were used between the fractural ends for strongly fixing the locking plate and keeping it attaching compactly against the bone laterally. In this series 23 nonunions were fixed with a locking compression plate (LCP) and 15 with LISS. A window of 3.5–4.5 mm diameter was opened at the nonunion site with an electric saw or electrical drill. Through this window debridement was performed to remove any sclerotic bone and fibrous tissue so as to get unobstructed access to the medullary canal. In 36 cases (94.7%) autogenous iliac bone was grafted into the defect and nonunion area, while in 2 cases mixed allogenic bone (Shanxi provincial tissue bank, China) was used. The surgical wound was then irrigated, negative pressure drainage placed under the deep fascia and the incision closed routinely.

Postoperative care

On the second postoperative day, patients were allowed to walk without weight bearing and to have their adjacent joints gently put through a passive assisted range of movement. Partial and full weight bearing was encouraged as early as possible, depending on the patient's general status and radiological signs of callus formation.

Definition of nonunion and union

Nonunion was defined by the presence of pain at the fracture site without radiographic evidence of healing 9 months after surgical fixation ¹² . Union was considered to have occurred when the patient was pain free on full weight bearing and there was radiological evidence of bridging callus across more than 75% of the fracture ends within six months ¹³ .

Evaluations and statistical analysis

All cases were reviewed at monthly intervals both clinically and radiographically if admitted. All relevant information regarding complications of treatment, joint function, discrepancy in limb length, rotational deformity, presence or absence of pain on full weight bearing, time for union, limb alignment and knee joint angular deformity was abstracted. Function of the hip, knee, ankle and shoulder was assessed with the Harris hip score ¹⁴ , hospital for special surgery (HSS) ¹⁵ , American Orthopaedic Foot and Ankle Society (AOFAS) ¹⁶ and Neer score systems ¹⁷ , respectively, and were ranked as excellent, good, fair and poor. All data were analyzed by a commercially available statistical database package (SPSS version 12.0, USA). Numeration data was described by frequency and measurement data by mean. A paired sample t‐test was used to determine difference in measurement data between pre‐operative and most recent time of postoperative follow‐up. The statistical significance level was set at α= 0.05.

Results

All patients were followed up regularly with a mean follow‐up of 11.6 months (range, 6–20 months). Union was achieved in all cases without further surgery, the average healing time being 5.3 months (range, 4–8 months). Superficial infections occurred immediately postoperatively in 3 cases (7.9%), these incisions healed with oral antibiotics and cleaning of the incision. Healing of the incision was delayed till four weeks in one case and the limb was 2 cm short in one case. There were no other complications, such as plate breakage, loose implants or nerve palsy, in this series. No patients had pain on full weight bearing at their most recent follow‐up. Three patients with humeral nonunions achieved union with excellent or good range of movement of their shoulder joints. Three patients with proximal femoral nonunion healed with excellent or good hip function. Among 26 nonunions around the knee, at the most recent follow‐up the HSS scores in 16 cases were over 85, 70–84 in 4, 60–69 in 5, and in only one case was less than 60 (this patient's bone defect was very severe and joint function very poor prior to surgery). These HSS scores, and the range of movement of the knee in the 26 cases of nonunion close to the knee, were significantly better than the pre‐operative findings (Table 1). Among the 6 nonunions of the distal tibia adjacent to the ankle, 4 cases were good or excellent and 2 were fair in regard to ankle function according to the AOFAS assessment. In summary, among 38 patients, 30 (78.9%) had excellent or good function, 7 (18.4%) fair and 1 (2.6%) poor in regard to the range of movement of joints. Two cases are shown in 1, 2.

Table 1.

Function and ranges of movement of 26 non‐union cases near the knee joint ()

Items	Pre‐operation	Latest follow‐up	t value	P value
Range of movement of the knee joint	60.1°± 4.7°	98.5°± 4.3°	3.86	<0.01
HSS scores	57.4 ± 6.5	83.6 ± 5.9	3.57	<0.01

Figure 1.

A 55‐year‐old man with left tibial and fibular fractures. (A) Preoperative radiograph showing that the intramedullary nail is loose, nonunion was diagnosed 12 months after the fractures; (B) Postoperative radiograph showing fixation with LCP combined with bone grafting; (C) Radiograph at 6 months post surgery showing that the nonunion has healed. Function was excellent by 12 months post surgery.

Figure 2.

A 34‐year‐old woman with fractures of the right distal humeral and olecranon fixed with an intramedullary nail. Nonunion was diagnosed after 15 months. (A) Preoperative radiograph showing the two distal locking screw nails are loose; (B) Postoperative radiograph showing fixation with LCP combined with bone grafting in the humerus and fixed with a hollow screw in the olecranon; (C) Radiograph 12 months post surgery showing the nonunion has healed, function is good.

Discussion

The main causes of long bone fracture eventuating in nonunion are instability of the fracture ends, insufficient bony contact and loss of blood supply ⁹ . All patients in this study had suffered intramedullary nailing failures due to intramedullary nail breakages in 17 cases (44.7%), loosening of screw nails in 19 (50%), and screw nail breakages in 2(5.3%). The site of nail breakage was the proximal of the two distal locking holes. Some factors that can increase the risk of nailing failure include unstable fracture patterns (segmental or comminuted), unsuitable nail size, technical errors in nail insertion, and early weight bearing ³ , ¹⁸ , ¹⁹ , ²⁰ . Reviewing the pathogenetic indicators, it has been reported that some patients become aware of nail breakage through experiencing sudden pain and instability at the fracture site without any obvious trauma being involved, suggesting that fatigue is another of the modes of failure ²¹ . The high percentage of nail breakage in the present study group indicates that surgical technique and choice of nail size would be the two main factors contributing to nonunion.

How to deal with non‐union after fracture fixation by intramedullary nailing? Several surgical techniques have been developed to treat these nonunions, including plate osteosynthesis, reamed intramedullary nailing, and external fixation with compression as well as locking plate. Exchanging the nail is one of the choices, and some studies have documented that excellent rates of union can then be achieved. Furlong et al. reported that 24 patients (96%) with femoral nonunion achieved union after exchanging the nail without any additional procedure ¹² . Megas et al. have reported achieving a 100% union rate within 6 months in 50 cases of nonunion of the tibia ⁴ . Therefore, many surgeons believed that exchanging the nail is a reliable method for treating nonunions which occur in the middle of long bone shafts in the lower extremity ¹² , ²² , ²³ .

External fixation is another possible choice for treatment of nonunion, and is the primary fixation method for nonunions with severe bone loss requiring distraction osteogenesis and bone transport, severe post‐traumatic scarring of the local soft tissues and chronic infection with viable bone ⁵ , ²⁴ . In addition, internal fixation with traditional plate and screw osteosynthesis has been abandoned because of instability and an increased risk of devascularization ²⁵ . Meanwhile, traditional implants carry a high risk of fixation failure when they are used to treat patients with osteoporosis or bone loss ²⁶ .

In order to ensure successful union, stable internal fixation with bone grafting is the current treatment of choice for nonunions ²⁷ , ²⁸ . Seventeen patients with nonunions located in the shafts of the long bones achieved union without complications in this study. Twenty‐one patients with nonunions located in the metaphysis also achieved union within the expected time, most of them gaining good or excellent results. The locking plate offers many advantages for recalcitrant cases, especially for fixation of metaphyseal fractures. Due to their design, the LISS being anatomically pre‐shaped to the adjacent joints of long bones, while the LCP is designed for fixation of the diaphyses of extremities, both of these devices allow the screws to lock to the plate, thereby creating a stable, fixed angle device which acts like an internal fixator ²⁹ . This means that compression forces and friction of the plate on the bone surface are not necessary to attain stability with the bone‐implant construct. This method of fixation preserves the local periosteal blood supply, especially when applied percutaneously. The locking plate, an internal fixator, combines the mechanical advantages of stable plate fixation with the biological concepts of external fixation ⁹ . Certainly, exchanging intramedullary nails to correct nonunions in the shafts of long bones may be a good choice; many reports have shown their efficacy and they have an obvious price advantage over locking plates. However, biomechanical analyses and clinical experiences have given rise to reports that intramedullary nailing can not achieve definite rotational stability at the fracture site ³⁰ , ³¹ , which would give rise to re‐failure. This potential failure should be avoided as far as possible for patients who have already experienced multiple surgical interventions. Ten cases (26.3%) had experienced more than two surgeries in our series.

In addition, bone grafting was applied through a window opened near the nonunion site. In this series, most cases (94.8%) were grafted with autogenous iliac bone; 2 cases (5.2%) with little bone loss were grafted with allogenic bone.

In this series, we used locking plate fixation combined with bone grafting to treat aseptic nonunion, and a 100% union rate was achieved within 6 months. Only 3 cases (7.9%) had superficial infection of the incision and these were healed by taking oral antibiotics combined with cleaning the incision. In addition one incision had delayed healing without any other complications. It is thought that autogenous cancellous or allogenic bone grafting is successful because of its well‐known osteoinductive and osteoconductive properties, as well as because of introduction of new osteogenic bone marrow cells to the fracture site ³⁰ , ³¹ , ³² , ³⁴ . Kettunen et al. have reported that, of 41 cases of tibial delayed unions or nonunions treated with percutaneous bone grafting, 37 (90.2%) achieved union within 13 weeks ³⁴ . Therefore, a locking plate combined with bone grafting is an effective treatment for non‐unions. In addition, 30 cases (78.9%) achieved excellent or good results and 7 (18.4%) fair, with only one case being poor in function when they were evaluated at the most recent follow‐up.

The present study is limited by its retrospective nature and the number of patients. Because we had no control group treated with alternative treatment procedures such as intramedullary nailing, we cannot comment on whether bone graft combined with locking plate is the better treatment for the management of long bone nonunion.

In conclusion, it has been shown that locking plate fixation is an effective treatment for long bone nonunions after intramedullary nailing fixation, especially in the area of the metaphysis. Locking plate fixation with bone grafting can provide reliable stability and promote bone union, has an acceptable complication rate and offers good quality of life.

Disclosure

The corresponding author, Chang‐qing Zhang, representing all authors, certifies that this research did not involve any financial and personal relationships with other people or organizations (such as employers, consultancies, stock ownership, honoraria, paid expert testimonies, patent applications/registrations, grants or other funding) that might pose potential conflicts of interest or bias the findings.