*Corresponding author. Biomechanics Laboratory, University atBu!alo, 162 Farber Hall, Bu!alo, NY 14214, USA. Tel.: #716-829-2982; fax: #716-829-3945.

E-mail address: medige@eng.bu!alo.edu (J. Medige)

Journal of Biomechanics 32 (1999) 861}864

Technical Note

E!ect of bone quality on the forces generated by compression screws

K.J. Faran!, N. Ichioka", M.A. Trzeciak", S. Han", J. Medige!,",*, O.J. Moy"!Department of Mechanical and Aerospace Engineering, State University of New York at Buwalo, Buwalo, NY, USA

"Department of Orthopaedic Surgery, State University of New York at Buwalo, Buwalo, NY, USA

Received 8 March 1999

Abstract

Internal "xation of the fractured scaphoid bone is used to promote union between bone fragments and to decrease wristimmobilization. Headless screws are commonly used because they minimize interference with articular surfaces and reduce tissueirritation and immobilization. In the present experiment, compressive force was measured as a function of bone quality for twoheadless screw types, the Herbert and the Acutrak. Forty-seven cylindrical samples of cancellous bone were prepared from fresh,previously frozen human cadaveric distal femora. The compressive forces generated as the screws were advanced into the specimenswere measured and correlated to the specimens' bone mineral density (BMD) and density. Over the range tested, the averagecompressive force for the Acutrack screw was approximately 42% higher than that of the Herbert. Statistical signi"cance, however,could not established because of the low statistical power of the test due to the inherent spread in the data. For the Acutrak screw,force was best "t to BMD and to density by second-order polynomials. Regression analysis indicated that similar correlations did notexist between force and BMD or between force and density for the Herbert screw. The correlation shown by the Acutrak screwindicates that it may be a more predictable as well as more e!ective system and therefore there may be some advantage in selecting thissystem. Furthermore, results suggest that the Acutrak screw generates greater forces with increasing bone density and could be moree!ective for a younger population. ( 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Internal fracture "xation; Bone quality; Cancellous bone; Bone screws

1. Introduction

Internal "xation of the fractured carpal scaphoid boneis used to promote union between bone fragments and todecrease wrist joint immobilization. Compression of thefracture site may be one of the more important considera-tions when selecting the appropriate hardware because itpromotes bone healing, provides stability and resistsrotation of the components. Although studies evaluatingthe e!ectiveness of available internal "xation systems inhealing scaphoid fractures (Carter et al., 1991; Herbertand Fisher, 1984; Herbert, 1986; Maudsley and Chen,1972; O'Brien and Herbert, 1985; Rankin et al., 1991;Shaw, 1987,1991; Sukul et al., 1990) have found thatheadless screws produce lower compressive forces, head-less screws are often favored because they minimize inter-

ference with articular surfaces and reduce tissue irritationand immobilization (Herbert, 1986; Shaw, 1987). In thisstudy, the compressive forces of two headless screw de-signs (Fig. 1) are compared. The Herbert (Zimmer Corp.,Warsaw, IN) system is a dual-pitched headless screwmade from titanium alloy which has been described ex-tensively (Herbert, 1985). The Acutrak (Acumed, Inc.,Beaverton, OR) system is made from titanium alloy andis a fully threaded, self-tapping, cannulated, and taperedheadless screw with varying thread pitch with both thediameter and pitch gradually increasing toward the trail-ing end. Two previous studies (Trzeciak et al., 1997; Tobyet al., 1997) compared the two designs but neither ac-counted for bone quality.

In the present experiment, compressive force is mea-sured as a function of bone mineral density (BMD) anddensity for the Herbert and the Acutrak designs. BMDis clinically measurable and is known to decrease inosteoporotic bone and this change may a!ect the perfor-mance of the screws. Compared to previous studies, thiswork ensured that the screw sizes were equivalent andmeasured forces generated in bone rather than synthetic

0021-9290/99/$ - see front matter ( 1999 Elsevier Science Ltd. All rights reserved.PII: S 0 0 2 1 - 9 2 9 0 ( 9 9 ) 0 0 0 7 6 - 7

Fig. 1. The two screw types tested. The Herbert screw (left) is a head-less, dual-pitched compression screw. The Acutrak system (right) isa headless, cannulated screw with a varying thread pitch. It is also fullythreaded, self-tapping, and tapered in pro"le.

Table 1Specimen properties and force measurements

Minimum Maximum Mean Std.Dev.

BMD (g/cm3) 0.06 0.259 0.156 0.053Acutrak Density (g/cm3) 0.94 1.35 1.13 0.10

Force (N) 0.45 9.02 3.20 2.26

BMD (g/cm3) 0.087 0.245 0.154 0.046Herbert Density (g/cm3) 1.02 1.33 1.14 0.09

Force (N) 0.30 5.08 2.26 1.25

material. One objective of the study is to determinewhether bone quality can a!ect the choice of the "xationdevice.

2. Methods

In compliance with guidelines established by the localhuman subjects ethical committee, 47 axial cylindricalsamples of cancellous bone, 15 mm in diameter and22 mm in length, were prepared from fresh, previouslyfrozen human cadaveric distal femora with mean age of70 yr (S.D."8). A band saw was used to cut slices per-pendicular to the long axis of the bone and a core drillwith 15 mm inner diameter was used to produce thecylindrical cross sections. The cylinder ends were "nishedusing a low-speed irrigated diamond saw.

Two measures of bone quality were used. The "rst,BMD, was determined by imaging the specimens witha quantitative computed tomography (QCT) scanner(GE9800, Minneapolis, MN) using 3.0 mm slices taken at3.0 mm intervals. The images were analyzed using imageanalysis software (C-MED, Virtual Visions, Cupertino,

CA) to convert pixel intensity to BMD units. The averagepixel intensity of at least six slices for each specimen wasconverted to a BMD value (g/cm3). The second assess-ment of bone quality involved the measurement of thedensity of each sample by dividing its entire (wet) mass byits volume. The samples were divided into two groupswhich had similar distributions of density and BMD(Table 1).

For mechanical testing, each sample was transectedperpendicular to its long axis using an irrigated diamondsaw at slow speed. The fragments were then reduced sothe samples could be pre-drilled, and tapped as neededfor the particular screw. A washer type force transducer(D/A656-01, Sensotec, Columbus, OH) was placed be-tween the two parts to measure the compressive forcegenerated as the screws were advanced into the specimen.Loads were recorded using data acquisition software(Asyst, Keithley Instruments, Inc., Cleveland, OH) andanalyzed using statistical software (Statview, AbacusConcepts, Inc., Berkeley, CA).

To compare the design features, the screw dimensionswere matched as closely as possible. The outer diameterof the Herbert screw was 3.0 mm; the diameter of thetapered Acutrak design ranged from 2.8 to 3.6 mm. Thelengths of the Herbert and Acutrak screws were 24 and25 mm, respectively.

Screws were advanced in small increments and thesystem was allowed to stabilize for approximately 5 sbetween increments to eliminate axial forces applied bythe investigator and to account for viscoelasticity (stressrelaxation) of the material. Initially, the `relaxeda forceincreased with each advance of the screw. The highestlevel of the stabilized force was interpreted as the max-imum compressive force sustainable with the screw. Itwas believed that this force would provide a more preciseprediction of the force that would remain after surgery.

3. Results

The mean value of the compressive force of the Acut-rak group was approximately 42% greater than thatof the Herbert group. Comparison t-tests showed no

862 K.J. Faran et al. / Journal of Biomechanics 32 (1999) 861}864

Fig. 2. Graph depicting the compressive forces generated by the Acut-rak and Herbert screws as a function of BMD. For the Acutrak screw,the relationship between force and BMD is best described with thequadratic expression, Force"3.10!0.0326(BMD)#0.000192(BMD)2(R2"0.58, p"0.0001). No relationship exists for the Herbert screw.

Fig. 3. Graph depicting the compressive forces generated by the Acut-rak and Herbert screws as a function of density. For the Acutraksystem, the relationship between force and density is best described bythe expression, Force"72.7!36.5(density)#65.8(density)2 (R2"0.60,p(0.0001). No relationship exists between force and density for theHerbert screw.

statistical di!erence (p"0.26) in average compressiveforce between the two screws (Table 1) but the statisticalpower of the test was only 0.22 due to the inherent spreadof the data.

At increasing densities, compressive forces increasedfor the Acutrak design. This rise in force resulted in anoverall "t of force to BMD and to density by secondorder polynomials (Figs. 2 and 3) for the Acutrak designand is consistent with studies showing that the compres-sive strength of bone increases with either density (Alhoet al., 1988; Leichter et al., 1992) or the square of density(Galante et al., 1970). Since BMD is closely related tobone density (Galante et al., 1970; Leichter, 1982) an

increase in density tends to re#ect greater presence ofminerals, which provide bone with most of its load carry-ing capabilities. The correlation coe$cient between forceand density (R2"0.60, p(0.0001) was slightly higherthan that between force and BMD (R2"0.58,p"0.0001), but the di!erence is not su$cient to elimin-ate BMD as an e!ective predictor of screw performancein the clinical setting.

No correlation between force and BMD (R2"

0.06, p"0.27 for linear; R2"0.09, p"0.34 for quad-ratic) or between force and density (R2"0.07,p"0.22 for linear; R2"0.10, p"0.33 for quadratic)exists for the Herbert screw (Fig. 3).

4. Discussion

It was hypothesized that di!erent screws may be moree!ective in generating compressive forces, based onBMD. One objective was to determine if bone qualitycould a!ect the selection of the screw used in fracture"xation. The present work is unique in that it considersbone quality as a factor in force generation.

This study assumes that using cadaveric bone insteadof a synthetic material will yield better estimates of clini-cal results. Although using scaphoids would have beenpreferable, samples were prepared from cadaveric femorabecause of material availability. By harvesting fromfemora, it was also possible to get wide ranges of BMDand density values among samples and to get more uni-form BMD values within each sample. Trzeciak et al.used freshly harvested scaphoids and determined statis-tically that the compressive force of the Acutrack wastwice that of the Herbert, but used only four sets of pairedscaphoids in their comparison. For a wide range of den-sity values and using cancellous bone, the present studyshowed a 42% greater average compressive force for theAcutrak screws over the Herbert. This result is clinicallysigni"cant although the low statistical power due tovariations in the specimens precluded statistical signi"-cance, as determined by the t-test.

Qualitatively, forces generated by the Acutrak designagree with the results of previous work. Rankin et al.(1991) did not measure bone density of their cadavericsamples, but noted that this property varied remarkablywithin the samples tested and that `softera bones wereobserved to have lower failure loads. Trzeciak et al.(1997) found that both screws produced greater compres-sive forces in higher density synthetic bone material(1.17 g/cm3) than in lower density material (0.78 g/cm3).

The pro"les of the two screws and the methods ofinsertion may contribute to the di!erences in e!ects ofBMD and density on generated compressive forces. TheHerbert screw is dual pitched and requires that the bonebe tapped prior to insertion. The smaller pitch on thetrailing end may strip the bone as the screw is inserted

K.J. Faran et al. / Journal of Biomechanics 32 (1999) 861}864 863

because of the pitch mismatch between the screw and thetapped hole. The Acutrak design has a tapered pro"lewith a varying pitch and is self-tapping. Although thevariable pitch may initiate bone failure as described forthe Herbert system, the tapered pro"le could be bene"-cial in minimizing this damage by maintaining greaterbone integrity. As the screw is inserted into the bone, theincreasing diameter continues to interact with untappedbone and there is decreased pitch mismatch.

Our results indicate that the Acutrak screw generatesgreater forces with increasing bone density. Although theaverage compressive forces of the screws are comparableover the low ends of the ranges tested, the forces gener-ated by the Acutrak screw show moderate correlation toboth bone density and BMD.This correlation providespredictability and may be a bene"t of selecting this sys-tem. Furthermore, results suggest that the Acutrak sys-tem could be more e!ective for a younger population,whose bones are more dense (Smith and Smith 1976).Since the two designs are approximately the same sizeand both are headless, other factors such as ease ofimplementation, cost, and resistance to bending alsoneed to be considered.

Acknowledgements

The authors wish to thank Acumed, Inc. and ZimmerCorp. for loans of the insertion systems.

References

Alho, A., Husby, T., Hoiseth, A., 1988. Bone mineral content andmechanical strength. An ex vivo study on human femora at autopsy.Clinical Orthopaedics and Related Research 227, 292}297.

Carter, F.M., Zimmerman, M.C., DiPaolo, D.M., Mackessy, R.P., Parsons, J.R., 1991. Biomechanical comparison of "xation devices inexperimental scaphoid osteotomies. The Journal of Hand Surgery 16A, 907}912.

Galante, J., Rostoker, W., Ray, R.D., 1970. Physical properties oftrabecular bone. Calci"ed Tissue Research 5, 236}246.

Herbert, T.J., Fisher, W.E., 1984. Management of the fracturedscaphoid using a new bone screw. The Journal of Bone and JointSurgery 66 B, 114}123.

Herbert, T.J., 1986. Use of the herbert bone screw in surgery of thewrist. Clinical Orthopaedics and Related Research 202, 81}92.

Leichter, I., Margulies, J.Y., Weinreb, A., Mizrahi, J., Robin, G.C.,Conforty, B., Makin, M., Bloch, B., 1992. The relationship bet-ween bone density, mineral content, and mechanical strength in thefemoral neck. Clinical Orthopaedics and Related Research 163,272}281.

Maudsley, R.H., Chen, S.C., 1972. Screw "xation in the management ofthe fractured carpal scaphoid. The Journal of Bone and Joint Surgery54 B(3), 432}441.

O'Brien, L., Herbert, T., 1985. Internal "xation of acute scaphoidfractures: a new approach to treatment. Australian and New ZealandJournal of Surgery 55, 387}389.

Rankin, G., Kuschner, S.H., Orlando, C., McKellop, H., Brien, W.W.,Sherman, R., 1991. A biomechanical evaluation of cannulated com-pressive screw for use in fractures of the scaphoid. The Journal ofHand Surgery 16 A, 1002}1010.

Shaw, J.A., 1987. A biomechanical comparison of scaphoid screws. TheJournal of Hand Surgery 12 A, 347}353.

Shaw, J.A., 1991. Biomechanical comparison of cannulated small bonescrews: a brief follow-up study. The Journal of Hand Surgery 16 A,998}1001.

Smith, C.B., Smith, D.A., 1976. Relations between age, mineral densityand mechanical properties of human femoral compacta. Acta Ortho-paedica Scandinavica 47, 496}502.

Sukul, D.M., Kaulesar, K.S., Johannes, E.J., Marti, R.K.,Klopper, P.J., 1990. Biomechanical measurements on scaphoidbone screws in an experimental model. Journal of Biomechanics23 (11), 1115}1121.

Toby, E.B., Butler, T.E., McCormack, T. J., Jayaraman, G., 1997.A comparison of "xation screws for the scaphoid during applicationof cyclical bending loads. The Journal of Bone and Joint Surgery(American Volume) 79-A(8), 1190-1197.

Trzeciak, M.A., Ichioka, N., Moy, O.J., Peimer, C.A., 1997. Bi-omechanical Comparison of Headless Screw Fixation Utilized inScaphoid Nonunion. In Proceedings of the American Fracture Asso-ciation (AFA), 59th Annual Meeting and Scienti"c Program. SanDiego, CA.

864 K.J. Faran et al. / Journal of Biomechanics 32 (1999) 861}864