Injury, Int. J. Care Injured 43 (2012) 1159–1165

Biomechanical analysis of second-generation headless compression screws

Soroush Assari a, Kurosh Darvish a, Asif M. Ilyas b,*a Department of Mechanical Engineering, Temple University, Philadelphia, PA 19122, United Statesb Rothman Institute, Department of Orthopaedic Surgery, Thomas Jefferson University, Philadelphia, PA 19107, United States

A R T I C L E I N F O

Article history:

Accepted 11 March 2012

Keywords:

Scaphoid fracture

Compression screws

Headless screws

Herbert-Whipple

Mini-Acutrak

Kompressor Mini

Twinfix

HCS

Sawbones

A B S T R A C T

Introduction: Headless Compression Screws (HCS) are commonly utilized for the fixation of small bone

and articular fractures. Recently several new second generation HCS (SG-HCS) have been introduced with

the purported benefits of improved biomechanical characteristics. We sought to determine and compare

the biomechanical efficiencies of these screws.

Material and methods: Five HCS including four second generation (Mini-Acutrak 2 (Acumed), Twinfix

(Stryker), Kompressor Mini (Integra), HCS 3.0 (Synthes)) and one first generation (Herbert-Whipple)

were studied. Polyurethane foam blocks that represented osteoporotic cancellous bone (0.16 g/cc) with a

simulated transverse fracture at the waist were utilized and five screws of each brand were tested for the

generated compression force and fastening torque during insertion with and without pre-drilling.

Results: The generated compression force was highest for Mini-Acutrak 2 (45.41 � 0.88 N) and lowest for

Herbert-Whipple (13.44 � 2.35 N) and forces of Twinfix, Kompressor Mini, HCS 3.0 were in between in

descending order. The compression force of SG-HCS increased slightly without pre-drilling but it was not

statistically significant while the fastening torque increased significantly. Slight over-fastening beyond the

recommended stage significantly reduced the compression force in Twinfix and Kompressor and had no or

moderate effect in other screws.

Conclusion: All SG-HCS demonstrated greater biomechanical characteristics than the first generation

Herbert-Whipple screw. The Mini-Acutrak 2 with a variable pitch design generated the maximum

compression force and showed the most reliability and sustainability. Screws with independently

rotating trailing heads (Twinfix and Kompressor Mini) demonstrated loss of compression with extra

turns. The increase of fastening torque due to over-fastening and loss of compression at the same time in

some screw designs, demonstrated how the fastening torque (applied by the surgeon) can be a

misleading measure of the compression force. Application of SG-HCS in osteoporotic bone without pre-

drilling can slightly increase the compression force.

� 2012 Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Injury

jo ur n al ho m epag e: ww w.els evier . c om / lo cat e/ in ju r y

Introduction

The Headless Compression Screw (HCS) was initially designedto internally fix displaced small bone fractures like scaphoidnonunions,1–3 capitelum4 and also osteotomies of hallux valguscorrection of tarsal bones (Fig. 1).5 Internal fixation options prior tothe HCS included open or percutaneous Kirschner-wire fixation oropen reduction with headed lag screws. K-wire fixation had apropensity for fracture distraction, fracture instability, andsecondary loss of reduction. Open lag screw fixation could resultin either poor compression or joint arthrosis as the head of thescrew would reside on the articular surface and therefore causesecondary joint injury.6

* Corresponding author at: Rothman Institute, Department of Orthopaedic

Surgery, Thomas Jefferson University, 925 Chestnut St, Philadelphia, PA 19107,

United States. Tel.: +1 610 755 3711; fax: +1 215 642 3633.

E-mail address: asif.ilyas@rothmaninstitute.com (A.M. Ilyas).

0020–1383/$ – see front matter � 2012 Elsevier Ltd. All rights reserved.

doi:10.1016/j.injury.2012.03.015

Herbert designed a single piece non-cannulated HCS in order toprovide internal compression and stability of the fracture whileavoiding any prominence of metal on the articular surface of thescaphoid or its joint space.7 HCS are favoured because they areembedded below the articular surface of the bone, which reducestissue irritation and immobilization. Whipple modified the Herbertscrew by developing a cannulated version to allow for more accuratescrew placement. During the 1990s, the use of cannulated HCS withplacement of a guidewire, open or percutaneously, from the volar anddorsal approach was popularized by several authors.8 Furthermore,the indication for use of HCS has grown to include the management ofminimally or non-displaced acute scaphoid fractures as well as to themanagement of other small bone or articular fractures.

Multiple manufactures are now offering newer or ‘‘secondgeneration’’ HCS (SG-HCS). These screws embody slight designmodifications that have been purported to improve biomechanicalcharacteristics. It is generally believed that compression plays animportant role in fracture stability by maintaining gap reductionand accelerating healing of cancellous bone.9–14 Consequently, the

Fig. 1. Applications for headless compression screws include small bone fractures and articular fractures such as (a) scaphoid and (b) capitellum fractures.

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–11651160

compression forces achieved by different bone screws and theirpullout strengths have been the focus of a few biomechanicalstudies. In this study, four frequently used SG-HCS, Mini-Acutrak 2screw (Acumed, Beaverton, OR), the Kompressor Mini screw(Integra, Plainsboro, NJ), the long thread HCS 3.0 (Synthes, Paoli,PA) and the Twinfix screw (Stryker, Mahwah, NJ) and a ‘‘firstgeneration’’ Herbert-Whipple screw (Zimmer, Warsaw, IN) wereinvestigated. Compression force of Mini-Acutrak (Acumed) hasbeen investigated by Adla et al.9 and Beadel et al.10 in syntheticbone and frozen scaphoid, respectively. Mini-Acutrak 2 is a newerproduct of Acumed after Mini-Acutrak to maintain the samecompression force with improved instrumentation. Hausmannet al.1 studied compression force of Twinfix using synthetic boneand Ramaswamy et al.19 investigated its pull-out force in threedifferent synthetic bone densities. Herbert-Whipple has beeninvestigated by Bailey et al.20 and Adla et al.9 in synthetic bones. Tothe knowledge of the authors, no study has been reported oncompression force of Kompressor Mini (Integra) and long threadHCS 3.0 (Synthes). So it was a clinical point of interest to design acomparative study to investigate these frequently used SG-HCS inone setup for their biomechanical performances, i.e., generatedcompression forces during insertion.

Furthermore, little attention has been paid to the applied torquemagnitude while inserting the screw. For a surgeon, the onlysensible reference of generated compression force is the appliedtorque which can be misleading. Additionally, the effect of pre-drilling on the maximum achievable compression force andapplied torque is another factor that has not been studied in theliterature. This study was designed to measure and compare thegenerated compression force and fastening torque during inser-tion, with and without pre-drilling, of four frequently used SG-HCScompared to the original Herbert-Whipple HCS.

Material and methods

Screws

Five cannulated HCS were studied including one ‘‘firstgeneration’’ Herbert-Whipple screw (Zimmer, Warsaw, IN) and

four SG-HCS including the: Mini-Acutrak 2 screw (Acumed,Beaverton, OR), the Kompressor Mini screw (Integra, Plainsboro,NJ), the long thread HCS 3.0 (Synthes, Paoli, PA) and the 3.2 mmTwinfix screw (Stryker, Mahwah, NJ) (Fig. 2). All the screws hadnominal major diameter of 3 mm and a length of 24 mm. Detaileddimensions of one sample of each screw type were measured usingan optical microscope at 10� magnification (Table 1). For eachscrew type, five samples were tested to account for any possiblemanufacturing errors.

Synthetic bone model

Previous studies have shown that the density and elasticmodulus of cancellous bone (e.g., scaphoid) are highly vari-able9,10,15–18 and this can affect the maximum achievable compres-sion force.9,16,19 Subsequently, achieving a meaningful statisticalsignificance would require a large number of cadaveric samples.Thus, in order to have uniform samples that allow the comparison of5 different screws, 50 identical bone models (10 for each screw type)were manufactured using solid rigid polyurethane foam (Sawbones,Vashon, WA), which has been established as an alternative testmedium for human cancellous bone.1,19–21 The American Society forTesting and Materials (ASTM F-1839–08) confirms that theuniformity and consistent properties of rigid polyurethane foammake it an ideal material for comparative testing of bones screwsand other medical devices and instruments. The chosen foamdensity (0.16 g/cc) with 1.6 MPa shear strength represents osteopo-rotic human cancellous bone19 and is used widely in previousstudies related to small bone fractures.19–21

Experimental setup

Fastening torque represents the surgeon’s physical sense ofresistance of purchase during screw placement. However, fasten-ing torque can be a misleading indication of interfragmentarycompression force. Therefore, the test setup was designed tomeasure both the fastening torque and compression forcesimultaneously. The 30 mm synthetic bone models were cutprecisely into two 15 mm thick pieces (CNC milling machine, HAAS

Fig. 2. The five headless compression screws studied included: Mini-Acutrak 2 screw (Acumed, Beaverton, OR), the HCS 3.0 (Synthes, Paoli, PA), the Herbert-Whipple screw

(Zimmer, Warsaw, IN), the Kompressor Mini screw (Integra, Plainsboro, NJ), and the Twinfix screw (Stryker, Mahwah, NJ).

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–1165 1161

Automation, Oxnard, CA). The setup was made such that both bonepieces could freely move along four rods transferring the fasteningtorque to a torque cell (model QWFK-8 M, Sensotec, Honeywell,Columbus, OH) (Fig. 3). A washer load cell (model LC8200,Omegadyne, Sunbury, OH) was placed between the two bonepieces to measure the compression force. An improvement in thissetup compared to previous studies was reduction of the gapbetween the bone pieces, which represents the fracture, from 3–6 mm to only 0.5 mm.9,15,16,20 This is important in investigation ofa fully threaded screw, like Mini-Acutrak 2, because with a largegap the threads located in the gap area are not engaged with bonepieces which may reduce the compression force.

Procedure

Both load cell and torque cell were connected to a dataacquisition system (SC-2311, National Instruments, Austin, TX)and their outputs were acquired during insertion of the screws andwere recorded at every quarter of turn. Zero revolution was definedas the stage in which screw head was flush with the bone surface.All the screws were applied according to their technical instruc-tions. However, in order to study the risk of loss of compressiondue to over-fastening beyond the recommended stage, fasteningcontinued until the compression force did not change significantly.Also the effect of pre-drilling on the maximum compression forceand fastening torque was evaluated since all the screws in thisstudy are self-cutting and in practice can be applied with orwithout pre-drilling. Five samples of each screw were tested withand without pre-drilling, resulting in a total of 50 tests.

Table 1Dimensions of investigated screws.

Screw Section Thread

Major dia. (mm) Minor dia. (mm) Depth

Mini-Acutrak 2 Leading 3.4–3.2 1.9 0.7

Trailing 3.2–3.5 1.9–3.2 0.7–0.2

Twinfix Leading 3.3 2.0 0.6

Trailing 4.0 3.2 0.4

Kompressor

Mini

Leading 2.1–2.7 1.4–2.0 0.4

Trailing 3.4–3.6 2.5–3.2 0.5–0.2

HCS 3.0 mm

Synthes

Leading 3.0 2.0 0.5

Trailing 3.5 3.0 0.25

Herbert-Whipple Leading 3.0 2.0 0.5

Trailing 3.8 2.4 0.7

Statistical method

All statistical analyses were performed using JMP 8.0.1 software(SAS Institute, Cary, NC). For each screw, the maximum compressionforce, the revolution stage at which the maximum compression wasreached and the fastening torque at that stage were compared withand without pre-drilling methods using Student’s t-Test (two sidedwith a = 0.05). Two-way analysis of variance was used to comparethe maximum compression force between screw types and theeffect of pre-drilling. All ranges given next to the average values anderror bars in graphs indicate standard error of the mean (SEM) unlessindicated (Some error bars are too small to be visible).

Results

Mini-Acutrak 2 (Acumed)

The compression force generated by Mini-Acutrak 2 and thefastening torque gradually increased with each turn until thecompression force reached 45.41 � 0.88 N (Fig. 4). The compressionforce did not change due to over-fastening. The maximum compres-sion force without pre-drilling (49.03 � 2.55 N) was higher but wasnot significantly different (p = 0.29). However, the fastening torque atthe same stage increased significantly by 87% (p < 0.0001) (Table 2).

Twinfix (Stryker)

Unlike the Mini-Acutrak, the maximum compression force byTwinfix (30.69 � 0.69 N) was reached between a quarter to a half

Mid-Shaft dia. (mm)

(mm) Pitch (mm) Length (mm) Numbers

1.25–1.1 24 22 NA

1.1–0.8

1.3 5.2 4 2.1

1.3 3.9 3

0.4 3.2 8 2.2

0.4 3.6 9

1.2 6.5 6 2.0

0.6 1.8 3

1.2 6.0 5 2.6

1.0 3.0 3

Fig. 3. The synthetic bone models were cut precisely into two 15 mm thick pieces with a torque cell and washer load cell placed in between with 0.5 mm gap simulating a

transvers fracture, to measure fastening torque and compression force, respectively. (a, b) Experimental setup mounted on a vibration isolation table. (c) Schematic of setup.

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–11651162

turn after unlocking the second stage of the screwdriver andthereafter only turning the trailing head. There was about 45% lossof compression force after a quarter to a half turn after unlocking thesecond stage. Without pre-drilling there was no significant (p = 0.086)increase in the maximum force, but a significant 73% increase(p < 0.0001) in the fastening torque. (Fig. 5).

Kompressor Mini (Integra)

The overall trend of having a distinct maximum compressionforce during the initial screw placement was also observed withthe Kompressor Mini, similar to Twinfix (Fig. 6). However, thecompression force gradually reached the maximum level of

Fig. 4. The compression and fastening torque of the Mini-Acut

20.5 � 1.18 N within one and a quarter turn following compressionof the trailing head alone. The loss of compression force due to a halfturn over-fastening was significantly (p = 0.02) less than Twinfix. Themaximum compression force without pre-drilling was 19% higher(p = 0.04). However, there was no significant increase in the fasteningtorque (p = 0.08). Unlike the Twinfix, after the maximum compressionforce was reached, the fastening torque was increasing while thecompression force was decreasing.

HCS 3.0 (Synthes)

Unlike the other screws, because of the different screw driverdesign of HCS 3.0, the zero revolution indicates the stage at which

rak 2 screw per revolution, with and without pre-drilling.

Table 2Comparison of screws at maximum compression force level.

Screws Pre-drilling Max. force

(N)

Rev. Torque

(Nmm)

Mini-Acutrak 2

(Acumed1)

With 45.4 � 0.9 1.40 � 0.17 114.0 � 2.7

W/O 49.0 � 2.5 0.90 � 0.10a 213.2 � 5.3a

Twinfix (Stryker1) With 30.7 � 0.7 0.50 � 0.00 42.1 � 3.5

W/O 34.4 � 1.7 0.35 � 0.06 72.8 � 2.1a

KompressorTM Mini With 20.8 � 1.2 1.15 � 0.10 37.0 � 2.3

W/O 24.3 � 0.7a 1.20 � 0.09 43.2 � 2.0

HCS 3.0 (Synthes1) With 17.3 � 1.0 1.75 � 0.00 27.8 � 2.5

W/O 17.7 � 2.4 1.75 � 0.00 46.3 � 3.7a

Herbert-Whipple

(Zimmer1)

With 13.4 � 2.4 1.05 � 0.17 31.0 � 2.2

W/O 10.1 � 0.4 0.95 � 0.35 71.0 � 2.9a

a Significantly different from ‘with pre-drilling’’ value.

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–1165 1163

the compression sleeve is held stationary and countersinking isstarted. As countersinking progresses, the compression forcereached 17.26 � 0.98 N (Fig. 7). The achievable compression forcewithout pre-drilling (17.65 � 0.98 N) was not significantly different(p = 0.797). However, there was a 67% increase in the fastening torque

Fig. 6. The compression and fastening torque of the Kompresso

Fig. 5. The compression and fastening torque of the Twinfi

Fig. 7. The compression and fastening torque of the Synthes HC

without pre-drilling (p = 0.007). Over-fastening of this screw aftercountersinking did not change the compression force significantly.

Herbert-Whipple (Zimmer)

Herbert-Whipple screw generated the least maximum com-pression force (13.44 � 2.35 N) among all five screws (Fig. 8) whichdid not change significantly (p = 0.23) without pre-drilling(10.10 � 0.39 N). Unlike other screws, the self-cutting flutes didnot work well while inserting this screw without pre-drilling and anadditional 20 N of axial force on the screw driver was required thatresulted in a 129% increase (p < 0.0001) in the fastening torque.

Comparison of screws

The values of maximum forces and the corresponding torquesand revolution stages for all screws are summarized in Fig. 9. Themaximum compression force was significantly different (p < 0.01)among all screws. The Mini-Acutrak 2 and Herbert-Whipplerespectively generated the most and the least compression forces.No pre-drilling had no significant effect on the average compression

r Mini screw per revolution, with and without pre-drilling.

x screw per revolution, with and without pre-drilling.

S 3.0 screw per revolution, with and without pre-drilling.

Fig. 8. The compression and fastening torque of the Herbert-Whipple screw per revolution, with and without pre-drilling.

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–11651164

force (p = 0.098) but increased the average fastening torque by 76%(p < 0.001). Although the compression force was generally reduceddue to pre-drilling, the interaction between the effect of pre-drillingand the type of screw was not significant (p = 0.07).

Discussion

The present study was performed to compare five commerciallyavailable and frequently used HCS. One was the original Herbert-Whipple screw and the other four represented the newer secondgeneration HCS (SG-HCS). In order to improve the comparison ofscrew performances and eliminate the variability in geometry andmechanical properties associated with cadaveric specimens, solidrigid polyurethane foam was used as a substitute to humanosteoporotic cancellous bone. A small gap between two boneblocks simulated a transverse fracture at the waist of small bones(e.g., scaphoid). This study design incorporated a number ofmodifications from previous studies. The gap between the twobone pieces was reduced from 3–6 mm in previous studies to0.5 mm which simulated fracture more realistically and allowedfor more screw threads to be engaged for fully threaded screws(e.g., Mini-Acutrak 2). A novel addition in the current design wasincorporating the measurement of the fastening torque which wasconsidered as the only sensible tactile reference for the surgeonwhile inserting the screw. This study showed that the torque canbe increasing, which may give the impression of more compressionforce, while in fact the compression force may be decreasing (e.g.,Kompressor Mini, Fig. 5).

The results showed that no pre-drilling slightly increased thegenerated compression force in SG-HCS (not statistically signifi-cant) and increased the fastening torque. In Kompressor Mini theincrease in the fastening torque was not significant, which can beattributed to its relatively small thread size. For inserting SG-HCSwithout pre-drilling, no extra axial force was required but forHerbert-Whipple an extra 20 N was required due to its tip design.

The Mini-Acutrak 2 generated the highest compression force(45.4 � 0.88 N) and showed no reduction due to over-fastening,

Fig. 9. Summary of maximum compression forces, with and without pre-drilling, for

each screw. All screws had significantly different compression force.

which can be attributed to its fully threaded, conical and variablepitch design. As this screw is inserted, the trailing part with largerminor diameter and smaller pitch becomes engaged causing thecompression force to gradually increase. These findings areconsistent with most of the reported results on Mini-Acutrak inthe literature which is an older product of Acumed substitutedrecently by Mini-Acutrak 2 with the goal of maintaining the samecompression force with improved surgical instrumentation. Adlaet al.9 reported a slightly higher compression force (50.37 SD 6.22 N),which can be a result of using a longer screw (26 mm) in simulatedcancellous bone with higher shear strength (3.5 MPa).9 In the studyof Beadel et al.10 on cadaveric scaphoid samples the reportedcompression force of Mini-Acutrak (93 � 56 N) was higher than theresult of this study (but not statistically significant due to the largestandard error), confirming the utilization of synthetic bone materialfor comparative biomechanical study.10 The pull-out force of Mini-Acutrak as reported by Baran et al.21 for the same foam material as inthis study was higher (67.21 SD 8.5 N) which is expected due tocomplete failure of the material engaged between the threads(destructive test).21

Twinfix and Kompressor Mini both have a distinct modulardesign with rotating trailing heads which generates most of thecompression. Twinfix had the second highest generated compres-sion force of 30.69 � 0.69 N, 31% more than Kompressor Mini whichcan be due to its larger trailing head, 4 mm for the Twinfix versus3.2 mm for the Kompressor Mini. However, Twinfix showed a higherreduction in compression force due to over-fastening than Kom-pressor Mini, which may be due to the fact that the rotating head ofTwinfix is axially fixed but in Kompressor Mini the trailing headtraverses axially as the screw is fastened. Therefore, as Twinfix screwis over-fastened it will damage the threads causing a significantirreversible reduction in compression force. There is no previousstudy on Kompressor Mini to the knowledge of authors. Ramaswamyet al.19 investigated the pull-out force of Twinfix in the same foamdensity of 0.16 g/cc and reported 234.87 N (52.8 lb) pull-out forcethat based on the screw cross-sectional area and threaded lengthcorresponds to at least 18.7 MPa compressive stress and 3.4 MPashear stress.19 Although the pull-out force is expected to be higherthan the generated compression force, based on the shear andcompression strength of the foam material (1.6 and 2.2 MParespectively) it is expected that the material would fail at muchlower loads and therefore these results should be considered withcaution.

The Herbert-Whipple had the least generated compressionforce of 13.4 � 2.4 N. Bailey et al.20 documented 16.7 N (95% CI of�6.7) as the maximum achievable compression force for Herbert-Whipple screw in the same foam material as in this study with a1 mm cortical foam layer with higher density of 0.64 g/cc at the top,which may explain their higher compression force.20 A cortical layerwas not considered in the current study since HCS are typicallyembedded about 3 mm below the cortical surface of the bone.

S. Assari et al. / Injury, Int. J. Care Injured 43 (2012) 1159–1165 1165

One of the limitations of this study was the comparison of thescrews’ performances in a specific synthetic bone density, whereasthe screws may show different behaviour in different bonedensities and porosities, particularly because their thread sizesare different. Moreover, synthetic bones have almost uniformspherical pores, while human cancellous bone has a complexthree-dimensional porosity pattern and this may affect themaximum achievable compression force. The current study didnot investigate the effect of fatigue (cyclic) loading. The bonedamage that may result due to fatigue loading may cause loss ofcompression and affect the performance of screws which demandsfurther investigation.

Conclusion

All second generation HCS (SG-HCS) demonstrated greaterbiomechanical characteristics than the first generation Herbert-Whipple screw. Mini-Acutrak2 generated the maximum compres-sion and showed no reduction due to over-fastening. Twinfix andKompressor Mini both demonstrated loss of compression withextra turns. HCS 3.0 and Herbert-Whipple screws demonstratedthe lowest compression. All SG-HCS resulted in increasedcompression force without pre-drilling that suggests no pre-drilling as a preferred approach for osteoporotic bones. Theincrease of fastening torque due to over-fastening and loss ofcompression at the same time in some screw designs, demon-strated how the fastening torque (applied by the surgeon) can be amisleading measure of the compression force.

Conflict of interest statement

None of the authors received any funding or financial support inthe production of this manuscript.

References

1. Hausmann J, Mayr W, Unger E, Benesch T, Vecsei V, Gabler C. Interfragmentarycompression forces of scaphoid screws in a sawbone cylinder model. Injury2007;38:763–8.

2. Fowler JR, Ilyas AM. Headless compression screw fixation of scaphoid fractures.Hand Clin 2010;26:351–61.

3. Haisman JM, Rohde RS, Weiland AJ. Acute fractures of the scaphoid. J Bone JointSurg 2006;88:2750.

4. Sano S, Rokkaku T, Saito S, Tokunaga S, Abe Y, Moriya H. Herbert screw fixationof capitellar fractures. J Shoulder Elbow Surg 2005;14:307–11.

5. Hofstaetter SG, Glisson RR, Alitz CJ, Trnka HJ, Easley ME. Biomechanical com-parison of screws and plates for hallux valgus opening-wedge and ludloffosteotomies. Clin Biomech 2008;23:101–8.

6. Burkhart KJ, Nowak TE, Appelmann P, Sternstein W, Rommens PM, Mueller LP.Screw fixation of radial head fractures: compression screw versus lag screw – abiomechanical comparison. Injury 2010;41:1015–9.

7. Herbert TJ, Fisher WE. Management of the fractured scaphoid using a new bonescrew. J Bone Joint Surg Br 1984;66:114.

8. Gutow AP. Percutaneous fixation of scaphoid fractures. J Am Acad Orthop Surg2007;15:474.

9. Adla DN, Kitsis C, Miles A. Compression forces generated by mini bone screws –a comparative study done on bone model. Injury 2005;36:65–70.

10. Beadel GP, Ferreira L, Johnson JA, King GJW. Interfragmentary compressionacross a simulated scaphoid fracture – analysis of 3 screws1. J Hand Surg2004;29:273–8.

11. Hart MB, Wu JJ, Chao E, Kelly P. External skeletal fixation of canine tibialosteotomies. Compression compared with no compression. J Bone Joint Surg1985;67:598.

12. Mayer G, Wolfs E. Animal experiments to examine the histology of fracturehealing in osteosynthesis with external fixation and compression. Arch OrthopTrauma Surg 1983;101:111–20.

13. Rosen H. Compression treatment of long bone pseudarthroses. Clin Orthop1979;138:154.

14. Sarmiento A, Schaeffer J, Beckerman L, Latta L, Enis J. Fracture healing in ratfemora as affected by functional weight-bearing. J Bone Joint Surg1977;59:369.

15. Lo I, King G, Milne A, Johnson J, Chess D. A biomechanical analysis of intrasca-phoid compression using the herbert scaphoid screw system: an in vitrocadaveric study. J Hand Surg J Br Soc Surg Hand 1998;23:209–13.

16. Lo I, King G, Patterson S, Johnson J, Chess D. A biomechanical analysis ofintrascaphoid compression using the 3.00 mm synthes cannulated screwand threaded washer: an in vitro cadaveric study. J Hand Surg (British andEuropean Volume) 2001;26:22.

17. Chapman J, Harrington R, Lee K, Anderson P, Tencer A, Kowalski D. Factorsaffecting the pullout strength of cancellous bone screws. J Biomech Eng1996;118:391.

18. Goldstein SA. The mechanical properties of trabecular bone: dependence onanatomic location and function. J Biomech 1987;20:1055–61.

19. Ramaswamy R, Evans S, Kosashvili Y. Holding power of variable pitch screws inosteoporotic, osteopenic and normal bone: are all screws created equal? Injury2010;41:179–83.

20. Bailey C, Kuiper J, Kelly C. Biomechanical evaluation of a new compositebioresorbable screw. J H Surg (British and European Volume) 2006;31:208.

21. Baran O, Sagol E, Oflaz H, Sarikanat M, Havitcioglu H. A biomechanical study onpreloaded compression effect on headless screws. Arch Orthop Trauma Surg2009;129:1601–5.