The Effect of timing of conversion from external fixation to secondary intramedullary nailing in experimental tibial fractures


  • The authors have no conflicts of interest and all participated in study design, animal surgery, data collection and analysis, and manuscript writing.


Diaphyseal tibial fractures with initial temporary external fixation (EF) are usually converted to intramedullary nailing (IMN) within 2 weeks, and no consensus on the optimal conversion time point exists. Current clinical practice is mainly based on estimation of the risk of postoperative infection. This is the first investigation of the effect of timing of such conversion on fracture healing. Forty male rats received a standardized tibial shaft osteotomy and EF. The animals were then randomly assigned to conversion to IMN at either 7 (group A, N = 10), 14 (group B, N = 10), or 30 (group C, N = 10) days after initial fixation. Group D (N = 10) served as a control group without conversion. Evaluation at 60 days included X-ray, DXA, and mechanical testing. Group A had significantly increased bone mineral content and callus area compared to the control group. Groups B and C showed significantly inferior mechanical bending strength and rigidity compared to both group A and the control group (D). The timing of the conversion procedure has a significant effect on fracture healing. Early conversion procedure did not improve healing compared to control, but was advantageous compared to late conversion (at 2 or 4 weeks) with higher mineralization and superior biomechanical properties. © 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 29:126–130, 2011

In severely traumatized patients who are poor candidates for an immediate intramedullary procedure, the initial treatment of the tibial fracture is often temporary external fixation (EF) in accordance with “the concept of damage control”.1–3 Also, in tibial fractures with severely traumatized soft tissue, a temporary EF is a treatment option until adequate soft tissue coverage is acquired.4 Often, when the patient recovers, there is a planned conversion to intramedullary nailing (IMN).5 The rationale for the early conversion is that IMN may provide enhanced bone healing over EF without the risk of pin tract infection.6, 7 However, advances in biomechanics and biomaterials have resulted in improvements of EF frames, and they can now remain in place for prolonged periods without degradation in the pin-bone surface.4 This calls for a clarification of the fracture segment biomechanical consequences of timing of such conversion and asks the question of when, if at all, the surgeon ought to perform a conversion more pressing.

Little has been published when it comes to theory or experiments on the topic of timing of such conversion procedures and the influence on biomechanical aspects of fracture healing. Most fracture healing studies focus on the initial fracture treatment or events late in the fracture repair process such as treatment of delayed union or nonunion.8–12 To the authors' knowledge, no study of the biomechanical effect of timing of the conversion procedure on fracture healing has been published.

With this background, we used an experimental tibial fracture model to study the effect of timing of conversion from initial EF to secondary IMN.


Animals and Surgical Procedures

Forty male rats (Wistar; Møllegårds Avlslaboratorium, Eiby, Denmark) were used in this study. The animals had a mean weight of 386 g and were housed in rodent cages with a lid with hinged water bottle divider and separate food area. Two rats in each cage received a standard rodent diet (RM3 (E); Special Diets Services, Witham, United Kingdom). The light cycles were 12/12 h. The experiment conformed to the Norwegian Council of Animal Research Code for the Care and Use of Animals for Experimental Purposes (project 37/2007).

Anesthesia (0.3 ml per 100 g body weight) was a working solution of fentanyl/fluanisone; Hypnorm®; Janssen Pharmaceutica, Beerse, Belgium) and midazolam (Dormicum®; Roche, Basel, Switzerland) administered subcutaneously before osteotomy and fracture surgery.13 The left tibia was exposed through a 20 mm anterior incision from the tibial tuberosity and in distal direction. The muscles on the medial and lateral aspect of the tibia were carefully elevated from the tibia and the anterior 2/3 of the bone was cut with a fine tooth circular saw blade mounted on an electric drill at the level of the anterior ridge. Then, the remaining 1/3 was manually broken, leaving the fibula intact. The fibula was left intact for rotational stability.14

The osteotomy was initially treated with an EF which has previously been described.15, 16 Core drill holes in the tibia were 0.8 mm. Four threaded steel pins (diameter = 1.0 mm) were inserted; two proximal to the fracture and two distal. Bone-fixator offset was 6 mm. The EF was placed anterolaterally providing the animal freedom of knee and ankle movement and apparent normal quadrupedal locomotion. Accurate fracture alignment and reduction was verified perioperatively both manually and visually. Careful soft tissue handling was performed leaving the medial and posterior muscle segment attached to the bone. The operation wound was closed in two layers with absorbable suture. Then, a layer of transparent film dressing was sprayed on the sutured wound as protection. The alignment and accurate reduction of the fracture was verified perioperatively both visually and manually. Buprenorphine (Temgesic®; Reckitt & Benckiser, Slough, UK), 0.05 mg per kilo body weight, was injected subcutaneously twice daily the first three postoperative days for analgesia.

The rats were randomly assigned to four different treatment groups. Groups A (N = 10), B (N = 10), and C (N = 10) which underwent conversion to IMN at days 7, 14, and 30, respectively, after initial treatment. Group D (control, N = 10) was left with EF for the total period of 60 days. During the conversion, we removed the EF carefully and inserted a 0.8 mm diameter nail from the proximal side into the bone marrow cavity by introducing the nail through the anterior tip of the tibial plateau, proceeding past the distal tibofibular junction, and with the knee in a flexed position. Then, the nail was cut plush to the bony surface at the insertion side. The animals were observed daily the first 3 days after both initial and conversion surgery and then on a weekly basis.

The rats were killed with an intraperitoneal injection of 100 mg per kilo pentobarbital (Mebumal®; Nycomed, Roskilde, Denmark) at 60 days after surgery. The tibias were dissected and examined visually and nails and external devices were carefully removed. Between dissection and radiological, densitometric, and mechanical examination, the bones were kept frozen at −80°C.

Bone Evaluations

X-ray images were taken on a standard clinical digital system (Siemens Axiom Aristos; Siemens AG; München; Germany). The X-ray tube settings were 46 kV, 1.0 mAs, and focus-to-film (source-to-image receptor) distance of 115 cm.

Dual-Energy Radiographic Absorptiometry (DXA) was performed using a densitometer system for research animals (Piximus; Lunar Corp., Madison, WI). The X-ray tube voltage was 80 kV with a current of 400 µA and a focal spot size of 0.25 mm × 0.25 mm. Focal spot-to-image receptor distance was 32 cm. The values for callus area (CA), bone mineral density (BMD), and bone mineral content (BMC) were automatically analyzed and calculated by the accompanying software from a 3.8 mm region of interest at the fracture site.

The tibias were placed in between gauze pads soaked with 0.9% saline before mechanical testing. A universal testing machine with servo-hydraulic mechanical linear drive actuator with 100 mm total vertical displacement and maximum axial tension loading capacity of 250 N (MTS 858 Mini Bionix; MTS Systems Corp., Eden Prairie, MN) tested every bone to failure with a 3-point cantilever bending test.17 Standard program settings included vertical travel speed of 160 mm/min. The log file was converted to a classic stress–strain curve and values for basic mechanical bone properties strength, rigidity, and work to fracture were obtained using a mathematical software package (Origin v 7.5; OriginLab Corp., Northampton, MA).18 Similar mechanical testing of cadaver tibia-implant constructions for EF (N = 5) and IMN (N = 5) showed initial mechanical bending rigidity of 3.2 ± 0.6 and 0.6 ± 0.1 N/mm, respectively (mean ± SE). In comparison, rigidity of intact tibia (N = 5) was 3.6 ± 0.3 N/mm.


Statistical data are expressed as mean and standard error of the mean. Based on results in a previous study on EF, a power analysis of the number of animals needed in our study was undertaken.19 With a one tailed test, level of significance at 0.05, and strength of 0.80, the number of animals needed in each group to discover an significant mean difference was 8. Due to possible surgical complications, we chose 10 animals in each of the four study groups. To test for statistical differences between the groups we used a one-way analysis of variance (ANOVA) and the Fischer LSD post hoc test when significant differences were indicated. The level of significance was set at a probability below 0.05.


Three rats, one in group B and two in group C, had extramedullary nails detected on X-ray and were excluded from the study. Radiographically, all tibias showed signs of healing with bridging of fracture line and the converted bones had little or no remains of the pin holes of the initial management. There were some periosteal callus formation in groups A, B, and C, but little in group D (Fig. 1).

Figure 1.

X-rays of rat tibia diaphyseal fractures 60 days after initial external fixation and conversion to secondary intramedullary nailing at (A) 7, (B) 14, and (C) 30 days. D: Represents the control group without conversion.

Group A had superior fracture mineralization (Table 1) compared to all the other groups in terms of DXA measured BMC. BMD tended to be higher in group A compared to control and was significantly increased compared to groups B and C. All converted groups (A, B, and C) presented increased bone area compared to the control group, but only the difference between group A and the control group was statistically significant in our study.

Table 1. Callus Area, Bone Mineral Density (BMD), and Bone Mineral Content (BMC) in a 3.8 mm Region of Interest at the Tibial Diaphyseal Fracture Site in Rat 60 Days after Osteotomy and Initial Stabilization With External Fixationa
 Group AGroup BGroup CGroup DOne-Way ANOVAb
  • Groups A, B, and C were converted to intramedullary nailing at days 7, 14, or 30, respectively.

  • a

    The values are mean ± SE.

  • b

    A LSD post hoc test of significant differences in mean was performed.

  • c

    Group A versus group D had p = 0.003.

  • d

    Group A versus both group B and C had p = 0.001.

  • e

    Group A versus group B had p = 0.001, versus group C p < 0.001, and versus group D p = 0.003.

Callus area (mm2)25 ± 0.723 ± 0.623 ± 1.020 ± 1.00.028c
BMD (10−3 g/cm2)22 ± 1.516 ± 2.317 ± 2.319 ± 2.80.002d
BMC (10−3 g)54 ± 4.338 ± 2.338 ± 1.939 ± 3.40.001e

In mechanical testing (Table 2), group A demonstrated superior mechanical properties as bending strength, rigidity, and work to fracture compared to the other two conversion groups B and C, but there were no significant differences in strength and rigidity between group A and the control group.

Table 2. Mechanical Strength, Rigidity, and Fracture Energy in Rat Tibial Diaphyseal Fractures 60 Days Postoperatively, Tested to Failure with a 3-Point Were Initially Stabilized with External Fixationa
 Group AGroup BGroup CGroup DOne-Way ANOVAb
  • Groups A, B, and C were converted to intramedullary nailing at days 7, 14, or 30, respectively.

  • a

    The values are mean ± SE.

  • b

    A post hoc LSD test was used to test for significant differences in mean.

  • c

    Group A versus B had p = 0.003 and versus C had p = 0.006. Group D versus B had p = 0.005 and versus C had p = 0.008.

  • d

    Group A versus B had p = 0.005 and versus C had p = 0.011. Group D versus B had p = 0.001 and versus C had p = 0.002.

  • e

    Group A versus group B had p = 0.006 and versus C had p = 0.010.

Strength (N)19 ± 2.810 ± 1.7110 ± 1.319 ± 2.30.002c
Rigidity (N/mm)4.0 ± 0.32.4 ± 0.22.6 ± 0.34.4 ± 0.60.001d
Fracture energy (N × mm)55 ± 1323 ± 5.726 ± 5.234 ± 4.60.026e


In our study, it is clearly demonstrated that the timing of the conversion procedure from initial temporary EF to secondary IMN have a significant impact on fracture healing. We found that continuing fracture treatment by EF until bony union gained strength and rigidity to the same degree as conversion to IMN 1 week after the fracture, while fracture energy was less, probably due to reduced CA and bone mineralization. However, delay in definitive nailing for 14 or 30 days impaired fracture healing significantly in this animal model.

Data obtained from animal studies cannot be used directly in human medicine. However, rats are widely used in experiments as they are quite similar to humans when it comes to the biology of bone and fracture healing.20, 21 Furthermore, the unilateral external fixator is frequently used in standard clinical orthopedics today,22 and mechanical testing to failure is still the “gold standard” of bone healing evaluation. For these reasons, our standardized experimental fracture model results may represent an interesting adjunct to clinical studies on these issues.

While it has become clinically evident that limited primary fracture stabilizing rather than definitive orthopedic procedures should be performed in patients with higher injury severity scores,23, 24 clinical decisions regarding the timing of secondary fracture surgery have been unclear. In one study, about 40% of trauma patients who underwent major secondary reconstructive surgery within 3 days after admission developed multiple organ failure,25 and some authors have delayed extensive orthopedic procedures until 72 h after injury. The issue of timing of definitive orthopedic procedures was investigated in a large survey of patients, and it was found that patients who developed multiple organ failure had secondary surgery between days 2 and 4, whereas patients without organ failure were operated between days 6 and 8 after initial injury.26 Our experimental set-up was based on these clinical observations.

Our experimental set-up included a standardized diaphyseal tibial fracture with limited soft tissue injury and an intact fibula. Thereby it differs from the clinical situation with extensive soft tissue damage and is more representative for patients with closed fractures, but in an immunological unstable situation where damage control orthopedics is an option. At 7 days, a soft tissue neocallus is formed.27 After nailing of the tibia, the fibula and the neocallus provide a relatively torsion-stable fracture segment, with flexibility in bending and axial movement.

Experimental studies have clearly demonstrated the potentially positive effect of less fixation rigidity and interfragmentary movements on bone healing.28 The IMN fracture fixation in our set-up has previously demonstrated higher degree of callus mineralization and biomechanical bone properties compared to EF.16 However, simply changing to a less rigid fracture fixation during healing does not necessarily enhance fracture healing. The observed advantage of early conversion compared to later conversion conforms to published results of other experiments where only early mechanical fracture segment stimuli have had a documented positive effect on bone healing.29

The CA intervention during the conversion procedure may partly explain why late conversion IMN led to an inferior bone healing process regarding mineralization and biomechanical properties. The concurrent soft tissue damage and its management is a documented predictor of the outcome and need for re-operation of a fracture patient.30, 31 The additional manipulation of the fractured leg represented by the conversion procedure may have interfered with the hardening and maturation of the fracture callus. More specifically, the physical removal of the four external fixator pins and insertion of an intramedullary nail through the fracture site may have influenced the biological fracture repair process in our experimental setup, as it would in a clinical situation. Furthermore, the fracture site penetration of the nail might have represented more of a stimulatory influence in the early neocallus in the early conversion group than in the more mature callus in groups B and C. In the late conversion groups, the conversion procedure, or the change in bone-implant fixation may have caused a delay in callus maturation and remodeling or a re-initiation of early repair processes with soft callus production, and this may partly explain the lower BMD and significant reduction in bending strength and rigidity observed in the late conversion groups compared to the control group and the early conversion group.

There exists an inherited risk of deep infection when a secondary conversion procedure in long bone fractures to IMN is executed. Musculoskeletal trauma is often complicated by a high risk for ischemia/reperfusion injuries and secondary infections and the risk increases with the degree of soft tissue injury and fracture wound contamination. Although favorable result have been reported when secondary nailing is delayed until after granulation of the pin sites,32 recent clinical studies indicate a significant increase of this risk in late conversion procedures, that is, after 28 days.6

On the basis of these clinical studies and the findings from our experiments, we suggest that the optimal biological window of time for the conversion procedure with a fracture healing enhancement potential and a low infection risk level may be within the first or second week after the initial fracture fixation. Otherwise, when the conversion to medullary nailing cannot be performed early due to the status of multiple injured patients or to the damaged soft tissue surrounding the fracture segment, maintaining the EF until bony union seems a viable option both in terms of biomechanical fracture healing and to avoid deep infection.

In conclusion, the results of our study indicate that the timing of the conversion from initial EF to IMN has a significant impact on bone healing. Our experiment supports the clinical practice of early conversion as soon as patient or local soft tissue condition permits with two important arguments. Firstly, early conversion to IMN induces an advantageous increase in mineralization and callus formation in the fracture segment. Secondly, late conversion to IMN significantly reduce the biomechanical properties of the callus segment compared with early conversion. In cases where early conversion cannot be performed, continuation of EF or conversion to a more stable frame configuration to allow patient ambulation and some weight bearing may be a treatment option as definitive fracture management.


Financial support for this study was provided by the University of Oslo, Faculty Division Akershus University Hospital and Institute of Surgical Research, Oslo University Hospital Rikshospitalet. The authors wish to thank Knut Rekdahl and the Mechanical Engineering Department, University of Oslo for skilled mechanical craftsmanship, engineer Per Ludvigsen at Oslo University Hospital Rikshospitalet and researcher Lise Sofie Nissen-Meyer, University of Oslo for technical assistance in mechanical testing and DXA measurements, radiographers Camilla Stolp and Helge Grindbakken at Oslo University Hospital Rikshospitalet for assistance in obtaining radiographs, and the staff at the Department of Comparative Medicine, Oslo University Hospital Rikshospitalet for housing our research animals.