British Journal of Oral and Maxillofacial Surgery 52 (2014) 850–853

Available online at www.sciencedirect.com

Biomechanical investigation of naso-orbitoethmoid traumaby finite element analysisHeike Huempfner-Hierl ∗, Andreas Schaller, Alexander Hemprich, Thomas HierlDepartment of Oral and Maxillofacial Plastic Surgery, Leipzig University, Liebigstrasse 12, 04103 Leipzig, Germany

Accepted 28 July 2014Available online 16 August 2014

Abstract

Naso-orbitoethmoid fractures account for 5% of all facial fractures. We used data derived from a white 34-year-old man to make a transientdynamic finite element model, which consisted of about 740 000 elements, to simulate fist-like impacts to this anatomically complex area.Finite element analysis showed a pattern of von Mises stresses beyond the yield criterion of bone that corresponded with fractures commonlyseen clinically. Finite element models can be used to simulate injuries to the human skull, and provide information about the pathogenesis ofdifferent types of fracture.© 2014 The British Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.

Keywords: naso-orbitoethmoid trauma; facial trauma; finite element analysis; biomechanics

Introduction

The naso-orbitoethmoid complex comprises the confluenceof the orbit, nose, maxilla, ethmoid sinuses, frontal bone, andfloor of the frontal sinus. Markowitz et al classified fracturesof the area into 3 types (I to III) according to the involvementof the medial canthal tendon.1 They can be part of a panfacialfracture, or localised. Kelley et al reported that they accountfor about 5% of all facial fractures and we have also foundthis in our experience.2

For many years, a large number of fractures of thenaso-orbitoethmoid complex were sustained in road trafficaccidents, but as cars have become safer they now occurless often.3 Nowadays, most facial fractures are caused by

∗ Corresponding author. Department of Oral and Maxillofacial PlasticSurgery, Leipzig University, Liebigstr. 12, 04103 Leipzig Germany.Tel.: +49 (0)341-9721163; fax: +49 (0)341-9721169.

E-mail address: heike.huempfner-hierl@medizin.uni-leipzig.de(H. Huempfner-Hierl).

interpersonal violence and sports accidents,4,5 and theseimpacts differ considerably in velocity and power from thosesustained in road crashes. Fractures caused by a single impactmight seem of minor clinical relevance compared with thosesustained in road accidents, which can be panfacial and asso-ciated with other life-threatening injuries, but the regionconsists of many anatomical structures and is close to partsof the anterior skull base where smaller fractures might havesevere consequences – for example, intense bleeding frominjury to the ethmoid vessels. We therefore examined thebiomechanical performance of the bones in the area when hitby single impacts, and the distribution, direction, and extentof the progress of stress in a transient dynamic finite elementanalysis.

Methods

We constructed a model of the midface to make a finiteelement analysis of fractures of the naso-orbitoethmoid

http://dx.doi.org/10.1016/j.bjoms.2014.07.2550266-4356/© 2014 The British Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.

H. Huempfner-Hierl et al. / British Journal of Oral and Maxillofacial Surgery 52 (2014) 850–853 851

Fig. 1. Impact by virtual impactors for study designs 1 and 2.

complex caused by a single impact. It consisted of 736 93410-node tetrahedrons. Data were derived from the computedtomography (CT) dataset of a healthy 34-year-old white man(Siemens Somatom Plus 4 Volume Zoom, 1 mm contiguousslices). After manual segmentation the dataset was exportedinto VRML (virtual reality modeling language), triangulated(VWorks 4.0® Cybermed, Korea), and imported into ANSYSICEM CFD 12.0.1® (ANSYS Inc, Canonsburg, USA).6 Toassign individual variables of the bony material to each ele-ment, we translated grey-scale values of the CT Hounsfieldscale into information on bone density, and used a BoneMatscript to calculate Young’s modulus for each element.7,8 Twodifferent study designs were chosen: the medial third of theinfraorbital rim (Fig. 1), and the junction between the nasalbone, maxillary nasal process, and lacrimal bone (Fig. 2).

To simulate the impact from one single fisticuff, a virtualbrass impactor (weight 412 g, density 8.4 g/cm3), Young’smodulus of 100 000 MPa, and Poisson’s ratio of 0.37 were

Fig. 2. Contact zones between impactor and bone for study designs 1 and 2.

Fig. 3. Von Mises stresses in study design 1 that correspond with paranasalfractures of the medial inferior orbital rim and fractures in the anterior orbitalfloor (scale is von Mises stresses in MPa).

modelled according to the experiments of Waterhouse et al.9

The velocity of impact was 6 m/second. We used a tran-sient mode of simulation as the interaction between the skulland impactor depended on time. The model was fixed at theoccipital condyles in all degrees of freedom. We assumed vonMises stresses of 150 MPa for the yield criterion of bone inthe skull.10 According to the regulations of our InstitutionalReview Board, approval was not needed for this investigation.

Results

In design 1, finite element analysis found a total impactof 7200 N for 1.3 msec. Von Mises stresses of more than150 MPa, which corresponded with fractures, were seen inthe medial inferior orbital rim paranasally, and in the anteriororbital floor (Fig. 3).

In design 2, finite element analysis showed a total impactof 6980 N for 2.6 msec. Von Mises stresses of more than150 MPa were seen in the medial orbital wall. The contralat-eral medial orbital wall was also affected, but here the yieldcriterion was not reached. High stresses even spread to theoccipital bone (Fig. 4).

In both designs minor stresses spread to the ipsilateral LeFort I plane, but did not involve the anterior base of the skullor the bony optical canal. The pattern of the areas where vonMises stresses exceeded the yield criterion is consistent withtypical fracture patterns seen in many patients (Figs. 5 and 6).

Discussion

When investigating the biomechanics of facial trauma it isdifficult to generate a practical and ethically acceptable studydesign that will deliver valid and reliable information. In thepast cadavers were often used. Nowadays, Le Fort’s studiesof 190111,12 would not be feasible for ethical reasons and

852 H. Huempfner-Hierl et al. / British Journal of Oral and Maxillofacial Surgery 52 (2014) 850–853

Fig. 4. Von Mises stresses in study design 2 that correspond with fracturesof the medial orbital rim (scale is von Mises stresses in MPa).

because of the lack of human cadavers. Post mortem alter-ations, and in most cases the age at death of the cadaver, whichoften is not that of a typical patient with a facial injury, lim-its the use of such studies. Elhammali et al found a mean(SD) age of 29.7 (12.8) years in 147 patients with sports-related maxillofacial fractures,5 and Roccia et al reported amean age of 28.5 years (range 11 – 72) in 138 patients withsports-related maxillofacial injuries (male:female ratio 8:1).4

Yamada and Evans found that the bony strength of cadavericspecimens with an age at death of 70 to 80 years was about20% to 30% lower than specimens from cadavers with an ageat death of 20 to 30 years.13 Our own model complied with thetypical age and sex of a patient with facial injuries.4,5 Animalexperiments, which use macaques and concern orthognathicquestions,14 are not appropriate because of ethical reasons,and because they cannot be transferred to the human anatomy.

Fig. 5. Computed tomogram of patient showing fracture pattern (arrows)similar to stresses in study design 1. The injury was caused by interpersonalviolence.

Fig. 6. Computed tomogram of patient with fracture pattern (arrows) similarto stresses in study design 2.

We therefore chose finite element analysis to generate validmodels for biomechanical tests.

Finite element analysis was introduced into biomechani-cal medical research in the 1990s. Takizawa et al reported anumerical computer simulation for the analysis of blowoutfractures in 1988,15 but their model was only 2-dimensionaland rather simple because of limited computing capacity, anddata were derived from dry human skulls. In 1996 Voo et alpresented their finite element models of the human head.16

They reported that the modelling of a biological structuresuch as a human head differs significantly from that of otherstructures. Nagasao et al published studies about blowoutfractures on 3-dimensional models based on finite elementanalysis in 2006.17 They presented a detailed 3-dimensionalmodel of about 240 000 finite elements, which allowed thor-ough analyses, but their CT-data were also derived from dryhuman skulls.

Previously, most published studies on finite element mod-els were based on cadaveric data, which have the advantagethat the experiments can be repeated, whereas those on gen-uine cadavers cannot. However, the statements concerningpost mortem alterations and the age of patients remain valid.

Our experiments differed from other investigations in sev-eral respects. Our model of the midface consisted of 736 93410-node tetrahedrons, which gave a high resolution. Unlikemany finite element analyses that use uniform biomechanicalvalues, we calculated individual biomechanical variables ofbone, which improved accuracy and realism. Peterson et alfound that different areas of bone in the dentate maxilla varyin thickness and material properties, and suggested that theseshould be incorporated into finite element models,18 as pre-viously done by Szwedowski et al.8

H. Huempfner-Hierl et al. / British Journal of Oral and Maxillofacial Surgery 52 (2014) 850–853 853

Regarding impact, we chose a transient simulation to ana-lyse the development of stresses within the model under timedependency as in an actual injury. Our model was designedto give more valid and reliable information than has beenreported in earlier studies because our data were derivedfrom a living person of a typical age, and because we cal-culated individual bony variables for each element and useda large number of elements. Inevitably there are limitations,and compared with studies on real models, finite elementanalysis has advantages and disadvantages. The main advan-tages are that experiments can be repeated, there are no ethicallimitations, and the study design can be changed and adjustedaccording to need. However, the model is not real and theremay be inaccuracies in segmentation, meshing, or variables inbiomechanical materials. Szwedowski et al compared finiteelement analysis with real models.8 His model, which wasequal to those in our study, accorded well with strain gaugemeasurements.

The predictive value of simulations has to be measuredwith their clinical correlations. We investigated commonfractures, but as it was the first study on this topic to ourknowledge, comparison with other work was not possible.

The simulation showed that forces resulting from one fist-like impact cause stresses beyond the yield criterion andaccord with typical fracture patterns commonly seen clini-cally for both designs. Stress did not spread to the anteriorbase of the skull or the bony optical canal, but this couldchange if the point of impact was in a different place and theforce was greater.

High resolution finite element models can simulate trau-matic insults to the human skull and should be used in furtherstudies. They can also provide information about the patho-genesis of different types of fracture.

Conflict of interest

We have no conflicts of interest.

References

1. Markowitz BL, Manson PN, Sargent L, et al. Management of the medialcanthal tendon in nasoethmoid orbital fractures: the importance of the

central fragment in classification and treatment. Plast Reconstr Surg1991;87:843–53.

2. Kelley P, Crawford M, Higuera S, Hollier LH. Two hundred ninety-fourconsecutive facial fractures in an urban trauma center: lessons learned.Plast Reconstr Surg 2005;116, 42e-9.

3. Murphy Jr RX, Birmingham KL, Okunski WJ, Wasser T. The influ-ence of airbag and restraining devices on the patterns of facialtrauma in motor vehicle collisions. Plast Reconstr Surg 2000;105:516–20.

4. Roccia F, Diaspro A, Nasi A, Berrone S. Management of sport-relatedmaxillofacial injuries. J Craniofac Surg 2008;19:377–82.

5. Elhammali N, Bremerich A, Rustemeyer J. Demographical andclinical aspects of sports-related maxillofacial and skull base frac-tures in hospitalized patients. Int J Oral Maxillofac Surg 2010;39:857–62.

6. Schaller A, Voigt C, Huempfner-Hierl H, Hemprich A, Hierl T. Tran-sient finite element analysis of a traumatic fracture of the zygomaticbone caused by a head collision. Int J Oral Maxillofac Surg 2012;41:66–73.

7. Taddei F, Pancanti A, Viceconti M. An improved method for the automaticmapping of computed tomography numbers onto finite element models.Med Eng Phys 2004;26:61–9.

8. Szwedowski TD, Whyne CM, Fialkov JA. Toward characterization ofcraniofacial biomechanics. J Craniofac Surg 2010;21:202–7.

9. Waterhouse N, Lyne J, Urdang M, Garey L. An investigation intothe mechanism of orbital blowout fractures. Br J Plast Surg 1999;52:607–12.

10. Nagasao T, Miyamoto J, Shimizu Y, Jiang H, Nakajima T. What happensbetween pure hydraulic and buckling mechanisms of blowout fractures?J Craniomaxillofac Surg 2010;38:306–13.

11. Le Fort R. Étude expérimentale sur les fractures de la machoiresupérieure. Revue de Chirurgie 1901;23:208–27.

12. Le Fort R. Étude expérimentale sur les fractures de la machoire supérieure[Experimental study on fractures of the upper jaw]. Revue de Chirurgie1901;23:360–79.

13. Yamada H, Evans FG. Ratios for age changes in the mechanical propertiesof human organs and tissues. In: Evans FG, editor. Strength of biologicalmaterials.. Baltimore: Williams and Wilkins; 1970. p. 255–80.

14. Nanda R, Goldin B. Biomechanical approaches to the study of alterationsof facial morphology. Am J Orthod 1980;78:213–26.

15. Takizawa H, Sugiura K, Baba M, et al. Structural mechanics of theblowout fracture: numerical computer simulation of orbital deformationby the finite element method. Neurosurgery 1988;22:1053–5.

16. Voo K, Kumaresan S, Pintar FA, Yoganandan N, Sances Jr A.Finite-element models of the human head. Med Biol Eng Comput1996;34:375–81.

17. Nagasao T, Miyamoto J, Nagasao M, et al. The effect of striking angleon the buckling mechanism in blowout fracture. Plast Reconstr Surg2006;117:2373–81.

18. Peterson J, Wang Q, Dechow PC. Material properties of the den-tate maxilla. Anat Rec A Discov Mol Cell Evol Biol 2006;288:962–72.