J. Biomedical Science and Engineering, 2010, 3, 448-453
doi:10.4236/jbise.2010.35062 Published Online May 2010 (http://www.SciRP.org/journal/jbise/
Published Online May 2010 in SciRes. http://www.scirp.org/journal/jbise
The effect of moment arm length on high angled femoral neck
fractures (Pauwels’ III)
Matthew S. LePine1, William R. Barfield1,3, John D. DesJardins2, Langdon A. Hartsock1
1Department of Orthopaedic Surgery, Medical University of South Carolina, Charleston, SC;
2Department of Bioengineering, Clemson University, Clemson, SC;
3Health & Human Performance, College of Charleston, Charleston, SC.
Email: firstname.lastname@example.org; email@example.com; firstname.lastname@example.org; email@example.com
Received 21 January 2010; revised 6 February 2010; accepted 8 February 2010.
This study investigated loads among five fixation
types (FT) [three cannulated screws (CS), dynamic
hip screws with and without derotational screws
(DHS-DS and DHS), and dynamic helical hip screws
with and without derotational screws (DHHS-DS and
DHHS)] across three fracture moment lengths (ML)
in Pauwels’ Type III fractures. Methods: Seventy-five
sawbones were tested (5 FT × 5 trials × 3 ML). The
study hypothesis was that significant differences in
axial loading to failure would be demonstrated when
CS was compared with the other four FT at the three
MLs. Each construct was exposed to an axial com-
pressive load to failure. Construct failure was defined
as 5 mm of migration at the fracture site or fixation
failure. Shapiro-Wilk was used to test for data nor-
mality. Subsequently, independent t-tests with Bon-
ferroni correction was used for paired comparisons.
Results: At fracture Moments A and B there were no
statistical differences between CS and the other FT.
At fracture Moment C all four FT yielded signifi-
cantly higher (p ≤ 0.001) loads compared with CS.
Conclusions: for basicervical fractures CS is a subop-
timal form of fixation compared with DHS and
DHHS both with and without derotation screws.
Keywords: Pauwels’ Fracture, Biomechanical Testing,
Pauwels’ Type III fracture is described as a fracture an-
gle of the femoral neck and shaft greater than 50º in the
coronal plane, [1-5] with dominating shear stress and
varus loading and less compressive loading and stress
[2,3]. As a consequence, internal fixation of these unsta-
ble fractures surgically has been associated with high
rates of non unions [3,6] sparking debate over the most
effective internal fixation technique [7-10]. Biomecha-
nical and clinical studies have shown noncomminuted
femoral neck fractures in the subcapital and transcervical
regions fixed with three cannulated screws yield optimal
stabilization results, [11-13] while fractures in the ba-
sicervical region show best results when fixed with a
sliding hip screw [13,14].
Other variables contribute to the nonunion rates of
fixed femoral neck fractures including; femoral neck
fracture type, vascular insufficiency, and inaccurate reduc-
tion . Stankewich et al. expanded this list to include
femoral head bone density, percent comminution of the
fracture surface, moment arm length, and orientation an-
gle of the fracture surface relative to the femoral shaft .
Although biomechanical tests have compared the ef-
fectiveness of various fixation devices for different frac-
tures,  no study has sought to make a direct com-
parison based on moment arm length while maintaining
a constant fracture angle.
The purpose of this study was to compare the failure
load of four different methods of fixation; Dynamic Hip
Screws with (DHS-DS) and without derotational screws
(DHS), and Dynamic Helical Hip Screw with (DHHS-
DS) and without derotational screws (DHHS) (Synthes
USA, West Chester, PA) for Pauwels’ Type III fracture
across three different fracture moment lengths with three
cannulated screws (CS), the traditional “gold standard”
[17-19]. The study hypothesis was that no statistically
significant differences in resistance to axial loading to
failure would be demonstrated when the CS was com-
pared with the other four fixation types at each of the
three moment lengths.
2. MATERIALS AND METHODS
Seventy-five new 3rd Generation Composite left femoral
sawbones (Sawbones Worldwide-Pacific Research Labo-
ratories, Inc. Vashon, WA) of equal density was utilized
for this experiment. The 3rd Generation Composite Bone
model is designed to simulate natural cortical bone th-
rough use of glass fibers and epoxy resin pressure in-
jected around a foam core. Density of the sawbones was
M. S. LePine et al. / J. Biomedical Science and Engineering 3 (2010) 448-453 449
Copyright © 2010 SciRes. JBiSE
1.7 g/cc with 90 MPa of tensile strength and a modulus
of 12.4 GPa. The compressive strength was 120 MPa
and the modulus was 7.6 GPa. Other investigators [20,21]
have found that 3rd Generation Composite femurs fall
within the normal range for bending, torsional and axial
loading compared with both fresh-frozen and dried re-
hydrated cadaver bone. Sawbones also show less vari-
ability compared with cadaveric specimens, with the
interfemur variability of composites being 20-200 times
lower than that for cadaveric specimens, thereby allow-
ing for smaller differences to be characterized with small
sample sizes [20,21]. Our apriori power analysis sup-
ported use of 5 samples/fixation/moment arm length.
Five sawbones for each fracture–fixation and moment
arm length were performed, resulting in a total of 75
constructs being tested (5 fixation types × 5 sawbo-
nes/fixation type × 3 moment arm lengths).
The three groups of sawbone fractures were created,
based on the fracture moment arm length (Figure 1; A = 2.0
cm, B = 3.1 cm, C = 4.2 cm). The moment arm distance
was measured from the femoral head center to the center of
the fracture line on the sawbone specimen using Vernier
calipers as shown in Figure 1. A custom cutting guide was
engineered to ensure fracture line reproducibility.
A femoral neck osteotomy at 70 to the horizontal
(Figure 1), was created by a single investigator with a
bandsaw in each of the sawbones, extending from the
superior femoral neck to subcapital, transcervical and
basicervical regions. All fractures were repaired by a
single investigator using one of five forms of fixation.
The three CSs (7.3 cannulated screw 100 mm, 16 mm
thread–Synthes, Paoli, PA, USA Product # 208.900)
were placed in an inverted triangle configuration at a
135. A cannulated screw set jig supplied by Synthes
was used for screw insertion to reduce variability.
The four other fixation types included: 1) DHS-DS (4
hole 135 Std plate-Synthes Product # 281.140, 90 mm
compression lag screw–Synthes Product # 280.900) with
a derotational screw (7.3 cannulated screw 100 mm, 16
mm thread–Synthes Product # 208.900); 2) DHS (4 hole
135 Std plate–Synthes Product # 281.140, 90mm
screw–Synthes Product # 280.900) without a derota-
tional screw; 3) DHHS (4 hole 135 Std LCP-Plate–
Synthes Product # 282.612, 100 mm helical screw–Sy-
nthes Product # 282.239) with a derotational screw (7.3
cannulated screw 100 mm, 16 mm thread–Synthes
Product # 208.900); 4) DHHS-DS (4 hole 135 std
LCP-Plate–Synthes Product # 282.612, 100 mm helical
Screw-Synthes Product # 282.239) without a derota-
tional screw. All hardware was inserted according to the
technique guide provide by the manufacturer. The 4-hole
plate has traditionally been used for fixation of proximal
femur fractures. A PMMA appliance (Figure 2) was
created which cupped the displaced femoral head and
proximal femoral neck to insure reduction accuracy. Im-
plants were properly positioned by direct visualization
on the surface of the bone. The hip lag screw guide wire
was consistently placed in the center of the femoral head
through use of an aiming device.
The angle of implant insertion was in accordance with
the technique guide provided by the manufacturer inclu-
ding pre-drilling of all holes prior to fixation (Figure 3).
A representative group of each fixation type was X-
rayed to insure anatomic reduction.
The femurs were uniformly potted in dental stone (M-
odern Materials) at 25 of adduction in the coronal plane
and neutral position in the sagittal plane to simulate the
terminal stance phase of gait [22-25]. The Instron Me-
chanical Testing System (Instron Bi-Axial Servo-Hy-
draulic Testing Machine Model # 8874-Norwood, MA)
incrementally loaded the femoral head with an evenly
applied compressive axial load through a polished stain-
less steel cup attached to the ram. Potted femurs were
securely locked in place with the appliance pictured in
Figure 4 for testing. The appliance was free to move in
the medio-lateral plane.
The loading protocol utilized was from Chang et al.,
1987 , designed to simulate single leg stance. In this
Figure 1. Schematic demonstration of mo-
ment arm length (A,B,C). A = Subcapital; B
= Transcervical; C = Basicervical.
Figure 2. PMMA appliance used to hold the
2-part Pauwels’ fracture in the vise for accurate
reduction of fractures.
450 M. S. LePine et al. / J. Biomedical Science and Engineering 3 (2010) 448-453
Copyright © 2010 SciRes.
Figure 3. Pictures of the five fixation forms.
Figure 4. Testing jig used for mechanical testing
single-leg phase, body weight is transferred ahead of the
forefoot, creating loading conditions of nearly five times
body weight . Constructs were positioned under a
servohydrolic materials testing system producing a ver-
tically-downward force. Displacement was applied at a
rate of 0.5 mm/s and terminated when the construct
reached failure criteria. Failure was defined as 5 mm of
migration at the fracture site through either translation
and/or a combination of migration and rotation of the
femoral head or catastrophic failure of device or saw-
Force data were captured using Instron’s Max v7.1
computer software at 100 Hz. Statistical analysis was
performed using SPSS Version 14 software. Shapiro-
Wilk was utilized to test for normality. Unpaired t-tests
were subsequently used to compare the four constructs;
DHS, DHS-DS, DHHS, and DHHS-DS to the CS con-
structs at each of the three moment lengths. Due to mul-
tiple unpaired comparisons a Bonferroni adjustment was
used. Our apriori alpha level was p ≤ 0.05. With four
comparisons the adjusted alpha was lowered to p ≤
Shapiro-Wilk test for normality indicated that the data
was normally distributed (> 0.05). At fracture Moment A
(subcapital) there was no statistical difference in yield
forces (p = 0.026) (Figure 5(a)). At fracture Moment B
(transcervical) there were no statistical differences in
yield forces (p = 0.022). (Figure 5(b)) At fracture Mo-
ment C (basicervical) all four fixation forms yielded
statistically significant (p ≤ 0.001) high axial loads when
compared with CS as seen in Figure 5(c).
In the present investigation five fixation types at three
different fracture moment lengths (A [subcapital], B
[transcervical], C [basicervical]), were axially loaded
and biomechanically testing in 75 sawbone models. Our
study hypothesis was that no statistically significant dif-
ferences in resistance to axial loading to failure would be
demonstrated when the CS was compared with the other
four fixation types at each of the three moment lengths.
Results supported our hypotheses at each moment length
except the basicervical model where statistically differ-
ent results were seen
Aminian et al. in the J Orthop Trauma in 2007, bio-
mechanically tested 32 cadaver specimens utilizing four
fixation types. Two of our fixation types were the same
as Aminian et al. (CS and DHS). Aminian et al. speci-
mens were cyclically loaded at 1400 N at 3 Hz for
10,000 cycles followed by a load-to failure test. All CS
specimens failed during incremental loading with a fail-
ure load of 862 N . Tan et al. tested five pairs of fresh
frozen cadaver femurs and used cyclical axial loading of
750 N at 0.5 Hz for 200 cycles  and found that the
load at the yield point was significantly higher in the
group with more horizontally oriented screws leading to
M. S. LePine et al. / J. Biomedical Science and Engineering 3 (2010) 448-453 451
Copyright © 2010 SciRes. JBiSE
Figure 5. Maximum axial force for fracture Moment
A** =statistical trend compared with CS.
the suggestion that two horizontal screws in the femoral
neck provide better stiffness.
In our study both the CS and DHS construct failed at
higher loads ([Moment A-3709.9 N] [Moment B-4423.4
N] [Moment C-3830.8 N]), which we partially attribute
to the lack of cyclical fatigue testing, specimen differ-
ences (cadaver bone vs. sawbones of uniform density)
and different mechanical testing based on our moment
arm length differences and different fixation types. The
moment arm distance, measured from the center of the
femoral head to the fracture line, plays a key role. As the
distance increases the torque across the fracture line in-
creases. The authors of the present work evaluated the
current English literature through traditional search
techniques and could not locate any studies that have
investigated the mechanical properties of various fixa-
tion techniques in Pauwels’ Fracture Type III with mo-
ment length considered.
Our testing protocol from Chang et al., 2002 , best
approximated the study goal of assessing maximal axial
compressive loading leading to eventual failure among
the fixation types, although other protocols exist [27-30].
The authors recognized through earlier unpublished pilot
data that there would be minimal flexure of the sawbone,
similar to what was observed in Aminian et al.
Our fracture moments results supported the null hy-
pothesis are at the subcapital and transcervical region,
but not at the basicervical region.
The fracture Moment C model yielded statistically
significant differences (p < 0.001) between CS and each
of the other four fixation types. Our results, based on the
use of synthetic femurs, suggest that for a subcapital
high angle fracture (Moment A) the DHHS and DHS-DS
may be better clinical choices for fixation than the stan-
dard CS and at the intermediate fracture Moment B the
DHHS-DS would be the preferred form of fixation al-
though statistically significant differences did not exist.
At fracture Moment C, a high angle basicervical fracture,
the study results suggest that all of the fixed angle fixa-
tion forms that we tested are superior to CS in combating
axial load. At Moment C, there is a shorter amount of
screw length supported by the intact femur and when a
force is applied the screw is not supported by adjacent
cortical bone due to the width at the base of the femoral
neck. Fixation with CS can lead to greater displacement
under the applied load.
The literature is inconclusive with respect to DHS fix-
ation. Baitner et al. demonstrated that the DHS construct
was significantly stronger than the cannulated screw
construct , however Husby et al., found no signifi-
cant differences between the same two constructs .
In a normal gait cycle at heel strike and toe off, the
hip experiences approximately 4-5 times body weight
. In the instance of a 100 kg (981 N) individual
walking, the hip joint can experience between 3924-
4905 N. These numbers approximate the values at fail-
ure found in our investigation. At fracture Moments A, B,
and C the loads for all fixation types are at or exceed the
upper range of axial loading compared with the example
provided for normal gait. Obviously, an in vitro study
and what occurs in vitro is much different due to soft
tissue damping which occurs in humans as we inde-
pendently ambulate. However, our biomechanical results
452 M. S. LePine et al. / J. Biomedical Science and Engineering 3 (2010) 448-453
Copyright © 2010 SciRes. JBiSE
suggest that fixation type is a function of the moment
arm length in Pauwels; Type III fracture.
The US population is getting older and progressively
more obese. A 120 kg (1177 N) patient can experience
4708-5885 N in the hip which surpasses the load bearing
capacity of several of the fixation forms tested inde-
pendent of moment arm length. In the event of a stumble,
forces in the hip joint approach eight times body weight
. The result of this could be catastrophic based on
the results demonstrated by this laboratory study.
The results found in this project are clinically relevant
because we have demonstrated that significant differ-
ences do exist across moment lengths and fixation types
in Pauwels Type III fractures. Even with adequate reduc-
tions, failure rates are higher with cannulated screws
than with fixed angle devices . Probe and Ward 
reported that both fixed angle and multiple screws with
divergent paths were superior to parallel screws. The
current work supports the prior body of knowledge asso-
ciated with the difficult clinical issue of surgical repair
of Pauwels’ Type III fractures.
Strengths of our study include: bone uniformity, frac-
ture accuracy, a single investigators placement of screws
and other types of fixation, and uniformity of testing and
recording of data. The weaknesses of our investigation
include uniaxial loading. One perceived weakness
among some readers may be that we used sawbones as
opposed to fresh cadaver femurs. While this perspective
has merit, our position is that by using a material that
possesses uniformity, our protocol measured the axial
loading of the constructs and was not dependent on hu-
man bone quality in terms of density. Varying bone den-
sities can introduce a confounding variable, which we
chose to control in our work. This position is supported
by studies reported in 1996 and 2001 using sawbones
and comparing the mechanical results with human ca-
daver bone. The findings in these studies support the
structural equivalent of composite bones with natural
bones in axial, bending and torsional loading [20,21].
Our loading protocol at failure approximated load during
single leg stance. Because of hip musculature stabiliza-
tion our fixation model does not completely replicate in
situ loading. Different directional force vectors can ne-
gatively affect fixation. Static loading in a uniaxial di-
rection does not mimic the complex tensile, compressive
and shear loads across the hip in situ.
This study is clinically relevant in that for subcapital
high angle fractures differences existed with the DHHS
and DHS-DS being better fixation choices when com-
pared with the CS construct. For transcervical fractures
the best choice was DHHS-DS and for high angle basi-
cervical fractures DHS, DHS-DS, DHHS, and DHHS-
DS were all axially superior to CS for fracture fixation.
However, as noted in a recent clinical paper that evalu-
ated the efficacy of internal fixation of Pauwels’ Type III
femoral neck fractures, despite advances clinically and
biomechanically the ideal fixation form remains unclear.
Clearly patient bone quality, patient age and implant
position all affect fixation results in patients as noted in a
recent clinical paper .
Finally, our model was used purely as a model for co-
mparison of stability of individual types of internal fixa-
tion. Our in vitro study results suggest that surgical fixa-
tion for Pauwels’ Type III unstable fractures is a fun-
ction of the moment arm distance and loading magni-
 Nilsson, L., Johansson, A. and Stromqvist, B. (1993)
Factors predicting healing complications in femoral neck
fractures: 138 patients followed for 2 years. Acta Ortho-
paedica Scandinavica, 64(2), 175-177.
 Parker, M.J. and Dynan, Y. (1998) Is Pauwels’ classifica-
tion still valid? Injury, 29(7), 521-523.
 Bartonicek, J. (2001) Pauwels’ classification of femoral
neck fractures: Correct interpretation of the original.
Journal of Orthopaedic Trauma, 15(5), 358-360.
 Pauwels, F. (1976) Biomechanics of the normal and dis-
eased hip. Springer, Berlin.
 Caviglia, H.A., Osorio, P.Q. and Comando D. (2002)
Classification and diagnosis of intracapsular fractures of
the proximal femur. Clinical Orthopaedics and Related
Research, 399(7), 17-27.
 Marti, R.K., Schuller, H.M. and Raaymakers, E.L. (1989)
Intertrochanteric osteotomy for non-union of the femoral
neck. Journal of Bone and Joint Surgery, 71(5), 782-787.
 Haidukewych, G.J., Liporace, F. and Gaines, R. (2005)
Pauwels’ type 3 vertical femoral neck fractures – What is
the best fixation device? Orthopaedic Trauma Association
21st Annual Meeting, 21(8), 2005, 130-131.
 Stankewich, C.J., Chapman, J., Muthusamy, R., Quaid,
G., Schemitsch, E., Tencer, A. and Ching, R.P. (1996)
Relationship of mechanical factors to the strength of
proximal femur fractures fixed with cancellous screws.
Journal of Orthopaedic Trauma, 10(4), 248-257.
 Weinrobe, M., Stankewich, C.J., Mueller, B., Tencer, A.F.
(1998) Predicting the mechanical outcome of femoral
neck fractures fixed with cancellous screw: An in vivo
study. Journal of Orthopaedic Trauma, 12(1), 27-37.
 Husby, T., Alho, A., Hoiseth, A. and Fonstelien, E. (1987)
Strength of femoral neck fixation, comparison of six
techniques in cadavers. Acta Orthopaedica Scandinavi-
ca, 58(6), 634-637.
 Barnes, R., Brown, J.T., Garden, R.S. and Nicoll, E.
(1976) Al subcapital fractures of the femur. Journal of
Bone and Joint Surgery, 58B(1), 2-24.
 Bout, C.A., Cannegieter, D.M. and Juttman, J.W. (1997)
Percutaneous cannulated screw fixation of femoral neck
fractures: The three point principle. Injury, 28(2), 135-139.
 Baitner, A.C., Maurer, S.G., Hickey, D.G., Jazrawi, L.M.,
Kummer, F.J., Goldman, S. and Koval, K.J. (1999) Ver-
tical shear fractures of the femoral neck. Clinical Ortho-
paedics & Related Research, 367(65), 300-305.
 Deneka, D.A., Simonian, P., Stankewich, E.D., Chapman,
J.R. and Tencer, A.F. (1997) Biomechanical comparison
M. S. LePine et al. / J. Biomedical Science and Engineering 3 (2010) 448-453 453
Copyright © 2010 SciRes. JBiSE
of internal fixation techniques for the treatment of unsta-
ble basicervical femoral neck fractures. Journal of Or-
thopaedic Trauma, 11(2), 337-343.
 LaVelle, D.G. (2003) Delayed union and nonunion of
fractures. In: Canale, S.T., Ed., Campbell’s Operative
Orthopaedics, St. Louis, (10)3, 3125-3165.
 Aminian, A., Gao, F., Fedoriw, W.W., Zhang, L.Q.,
Kalainov, D.M. and Mark, B.R. (2007) Vertically orient-
ed femoral neck fractures:mechanical analysis of four fi-
xation techniques. Journal of Orthopaedic Trauma, 21(8),
 Hernefaulk, L. and Messner, K. (1995) Femoral stiffness
after osteosynthesis of a subcapital osteotomy in osteo-
porotic bone: An in vitro comparison of nine fixation
methods. Journal of Bone and Joint Surgery, 9(58B), 2-
 Kuokkanen, H., Korkala, O., Antti-Poika, I., Tolonen, J.,
Lehtimaki, M.Y. and Silvennoinen, T. (1991) Three can-
cellous bone screws versus a screw-angle plate in the
treatment of Garden I & II fractures of the femoral neck.
Acta Orthopaedica Belgium, 57(1), 53-57.
 Parker, M.J. and Blundell, C. (1998) Choice of implant
for internal fixation of femoral neck fractures: Meta-
analysis of 25 randomized trials including 4,925 patients.
Acta Orthopaedica Scandinavica, 69(2), 138-143.
 Christofolini, L., Vineconti, M., Cappello, A. and Toni, A.
(1996) Mechanical validation of whole bone composite
femur models. Journal of Biomechanics, 29(4), 525-535.
 Heiner, A.D. and Brown, T.D. (2001) Structural proper-
ties of a new design of composite replicate femurs and
tibias. Journal of Biomechanics, 34(6), 773-781.
 Choueka, J., Koval, K.J., Kummer, F.J., Crawford, G. and
Zuckerman, J.D. (1995) Biomechanical comparison of
the sliding hip screw and the dome plunger: Effects of
material and fixation design. Journal of Bone and Joint
Surgery, 77-B(2), 277-283.
 Goh, J.C.H., Shah, K.M. and Bose, K. (1995) Biome-
chanical study on femoral neck fracture fixation in rela-
tion to bone mineral density. Clinical Biomechanics,
 Kubiak, E.N., Bong, M., Park, S.S., Kummer, F., Egol, K.
and Koval, K.J. (2004) Intramedullary fixation of unsta-
ble intertrochanteric hip fractures: One or two lag screws.
Journal of Orthopaedic Trauma, 18(1), 12-17.
 Chang, W.S., Zuckerman, J.D., Kummer, F.J. and Frankel,
V.H. (1987) Biomechanical evaluation of anatomic re-
duction versus medial displacement osteotomy in unsta-
ble intertrochanteric fractures. Clinical Orthopaedics &
Related Research, 225(6), 141-146.
 (1986) Gait analysis. In: McAinsh, T.F., Ed., Physics in
Medicine & Biology Encyclopedia: Medical Physics,
Bioengineering and Biophysics, Pergamon Press Inc,
New York, 1(5), 356.
 Tan, V., Wong, K.L., Born, C.T, Harten, R. and DeLong,
W.G. (2007) Two-screw femoral neck fracture fixation: A
biomechanical analysis of 2 different confirurations.
American Journal of Orthopedics, 36(9), 481-485.
 Chaim, S.H., Mukherjee, D.P., Ogden, A.L., Mayeux, R.
H., Sadasivan, K.K. and Albright, J.A. (2002) A biome-
chanical study of femoral neck fracture fixation with the
VHS-Vari-Angle hip fixation system. American Journal
of Orthopedics, 31(Suppl 1), 22-24.
 Aboulafia, A.J., Price, M.M., Kennon, R.E. and Hutton,
W.C. (1999) A comparison of mechanical strength of the
femoral neck following locked intramedullary nailing
using oblique versus transverse proximal screws. Journal
of Orthopaedic Trauma, 13(3), 160-163.
 Kukla, C., Gaebler, C., Pichl, R.W., Prokesch, R., Heinze,
G. and Heinz, T. (2002) Predictive geometric factors in a
standardized model of femoral neck fractures. Experi-
mental study of cadaveric human femurs. Injury, 33,
 Probe, R. and Ward, R. (2006) Internal fixation of femo-
ral neck fractures. Journal of the Academy of Orthopae-
dic Surgeons, 14(9), 565-571.
 Liporace, F., Gaines, R., Collinge, C. and Haidukewych,
G.J. (2008) Results of internal fixation of Pauwels’ type-3
vertical femoral neck fractures. Journal of Bone and
Joint Surgery, 90(A), 1654-1659.