In children fractures of the femoral diaphysis occur in around 2-3% of all fractures and is the second most frequent localization affecting the lower extremity. The guidelines of the German Society of Pediatric Surgery recommend elastic stable intramedullary nailing (ESIN) for displaced fractures even in complex or spiral fractures for children older than 2- to 3-years-of-age [1]. This is based upon reported rapid recovery, fast reintegration of the patients, a reduction of the possible negative effects of immobilization and few to no complications. Further on, elastic stable nails are preferred to avoid any possible damage to the open physis and the possible complication of avascular necrosis of the femoral head following the implantation of a rigid nail [2–7]. The technique mostly used is the so called 2-C-shaped configuration, in which two nails are implanted in an ascendig matter from the distal metaphysis of the femur [8–11]. On the other hand, skin problems and soft tissue irritation are well-known problems and some publications reported complications in up to 50% of cases with an increased tendency towards complex fracture types and older children weighting more than 40 kg [12–17]. As a consequence of this type of problem, other authors prefer submuscular plating [18, 19] or external fixation for those instable fractures [20]. Rapp and co-workers analyzed 31 children with femoral shaft fractures treated by ESIN-osteosynthesis: 4 needed an additional cast and 8 patients required reoperation. Indications were 4 varus deformities and 4 shortenings of the fracture (“telescoping”). In 3 cases these were caused by technical errors during implantation and 5 fractures which could not sufficiently stabilized owing to technical limitations of the method itself [21]. In conclusion, shortening, recurvation and varus seem to be the most important clinical problems associated with ESIN in femoral fractures.

End caps were recommended to improve stability of the ESIN-osteosynthesis [22], but this could not be proved in our biomechanical in-vitro testing [23]. In the current study we aimed to analyze the effect of a 3rd ESIN on stiffness, which is a valid stability parameter in biomechanical testing [24, 25]. Therefore in our spiral femur fracture model we compared osteosynthesis with 2C-shaped nails (2C) vs. 2C-shaped nails with the 3rd from antero-medial (3CM) and 2C-shaped nails with the 3rd from antero-lateral (3CL), respectively. If an improvement in stiffness through one or both 3 nail configurations (3CM/3CL) occurred, these modifications would be evaluated in children with complex femur fractures.

Elastic stable intramedullary nailing (ESIN) for displaced pediatric femoral fractures has become the treatment of choice for children [1]. Nevertheless complications of the ESIN-osteosynthesis occur due to technical errors during application and limitations of the method itself [12–17, 30–32]. As a conclusion of literature and own data, most frequent postoperative complications seem to be shortening (“telescoping”), recurvation and varus deformation resulting in postoperative misalignment, re-do operations and the additional application of a cast or external fixation [33, 34]. As standardized biomechanical studies are known to be an important precursor to clinical studies, we were sure that further experiments in this area would increase our understanding of ESIN-osteosynthesis. Thus, an adolescent-sized femur model with a standardized spiral fracture was used, because the primary targets were complex fractures in older children or weighing more than 40 kg. Because of the complexity of long spiral fractures, testing of all relevant variables was performed: 4-point bending (corresponding to deformation forces as antecurvation, recurvation, valgus and varus), external and internal rotation as well as axial 0° and physiological 9° compression (corresponding to shortening or “telescoping”).

Few reports of further biomechanical testing of the ESIN-osteosynthesis could be found; most of these where not standardized, depicted a small number of cases [35, 36], evaluated only a few selected planes [35, 37, 38] or studied exclusively transverse and oblique fractures [36]. Benz and co-workers investigated changes of stability during 4-point-bending and torsion, but they only achieved sufficient 2-nail flexible intramedullary fixation for testing in 3 out of 10 canine bones [39].

Through our biomechanical model we compared the “classical” 2C configuration (2C-shaped 3.5 mm titanium nails) with 2 modified versions, each with an additional 2.5 mm titanium nail placed either from antero-medial (3CM) or antero-lateral (3CL). All the specimens with 3 ESIN achieved a better alignment and macroscopic stiffness of the fracture. In our specific spiral fracture type with rotation from distal lateral to cranial medial, the 3rd nail from anterior-lateral was effective in reducing 2 of the 3 causes of poor clinical outcomes mentioned above: the danger of shortening (better results in 9° compression) and recurvation (better results in anterior-posterior). That position of the 3rd intramedullary nail provided more stability for this specific type of the fracture which also might explain why contradictory results occurred in posterior-anterior versus anterior-posterior and internal versus external rotation. The further clinical problem, the valgus deviation, was not influenced by either modified configuration.

Nevertheless variations in the course of the 3rd nail might influence this system: If the planes between the 3rd nail and each of the other 2 were almost parallel, the increase in stability would barely be noticeable. On the other hand, positioning the 3rd nail in an almost equilateral triangle with the first 2 nails could result in significantly increased stability. Further CT-imaging studies with our 3 nail specimen will test this hypothesis.

A limitation of this first biomechanical study on the use of 3 nails is that a synthetic bone model was used, which cannot precisely reproduce all in-vivo conditions. However, because of the high repeatability due to minimal inter-individual variability [40–42], it was used successfully in previous biomechanical studies [23, 26, 27]. Furthermore, due to the configuration of the synthetic bone model, the nails could not be placed as proximally as is ideal in real operations. On the other hand, the planned configuration of the nails can be achieved more precisely in an experimental setup with an identical surgical technique, an exact pre-bending and an even introduction of the nails, than in a real surgical situation.

In spite of these limitations of the biomechanical study, our results were transferred successfully to a clinical setting. Osteosynthesis with 3 nails was first performed by a pediatric surgeon, specialized in pediatric traumatology (MMK) and then transferred to other members of the pediatric surgery department. Although the 3CL-configuration was initially preferred, the position of the 3rd nail was later chosen in accordance with the plane which was assumed to be the least stable. No clinically relevant deviation of axis or shortening was documented in osteosynthesis with 3 nails in 16 patients. None of the fractures required additional stabilization or re-operation. In 2 patients the 3 ESIN-osteosynthesis did not gain enough stability: Because of a contra lateral fracture of the tibia, a modification with 4 nails was chosen to reach early mobilization with weight bearing (No. 7) and in one long spiral fracture an angle-stable osteosynthesis had to be performed (No. 18). All patients reached the maximum Harris score of 100 points in the follow-up evaluations.

In essence, the fixation of complex femur fractures remains technically demanding and this comparison shows the high relevance of pre-clinical experimental studies. A high success rate can be achieved using insights gained from in-vitro studies, thus preventing children and adolescents from undergoing unnecessary clinical trials. We can recommend the use of a 3 ESIN-osteosynthesis in pediatric femur fractures, as long as the biomechanical characteristics of the fracture type are taken into account in the placement of the 3rd nail.

In this biomechanical model the 3rd nail from antero-lateral ‘3CL’ was effective in reducing two of the three causes of poor clinical outcomes in ESIN-osteosynthesis: the danger of shortening (significantly higher stiffness in 9° compression) and recurvation (significantly higher stiffness in anterior-posterior bending). These in-vitro results were successfully transmitted into clinical practice and adapted (3rd nail from antero-lateral or -medial) to the individual character of each femur shaft fracture. Through the successful transmission of our biomechanical results into a clinical setting we are able to recommend the use of a 3rd nail in ESIN in pediatric femur fractures to improve the stability of the osteosynthesis and to reduce peri- and postoperative complications. As a consequence, at our hospital all femoral diaphyseal fractures in children/adolescents with open physis are treated with this modified ESIN-technique. However, a perfect technique while implanting the first two nails is always necessary.