SCIENTIFIC ARTICLE

Comparison of Screw Trajectory on Stability of

Oblique Scaphoid Fractures: A Mechanical Study
Gregory K. Faucher, MD, M. Leslie Golden III, MS, Kyle R. Sweeney, MD,
William C. Hutton, PhD, Claudius D. Jarrett, MD

Purpose To determine whether a screw placed perpendicular to the fracture line in an oblique
scaphoid fracture will provide fixation strength that is comparable with that of a centrally
placed screw.
Methods Oblique osteotomies were made along the dorsal sulcus of 8 matched pairs of
cadaveric scaphoids. One scaphoid from each pair was randomized to receive a screw placed
centrally down the long axis. In the other scaphoid, a screw was placed perpendicular to the
osteotomy. Each scaphoid underwent cyclic loading from 80 N to 120 N at 1 Hz. Cyclic
loading was carried out until 2 mm of fracture displacement occurred or 4,000 cycles was
reached. The specimens that reached the 4,000-cycle limit were then loaded to failure. Screw
length, number of cycles, and load to failure were compared between the groups.
Results We found no difference in number of cycles or load to failure between the 2 groups.
Screws placed perpendicular to the fracture line were significantly shorter than screws placed
down the central axis.
Conclusions A perpendicularly placed screw provides equivalent strength to one placed along
the central axis.
Clinical relevance Compared with a screw placed centrally in an oblique scaphoid fracture, a
screw placed perpendicular to the fracture line allows the use of a shorter screw without
sacrificing strength of fixation. (J Hand Surg Am. 2014;39(3):430e435. Copyright Ó 2014 by
the American Society for Surgery of the Hand. All rights reserved.)
Key words Biomechanics, internal fixation, scaphoid fracture, wrist.

H
acute scaphoid
ERBERT AND FISHER CLASSIFIED
waist fractures into 2 main groups: simple
transverse waist fractures (type B2) and
distal oblique fractures (type B1). Oblique scaphoid
fractures are difficult to treat owing to the shear
From the Hand and Upper Extremity Surgery, The Emory Orthopaedic Center, Department
of Orthopaedic Surgery, Emory University School of Medicine, Atlanta, GA.
Received for publication April 19, 2013; accepted in revised form December 11, 2013.
No benefits in any form have been received or will be received related directly or
indirectly to the subject of this article.
Corresponding author: Claudius D. Jarrett, MD, Hand and Upper Extremity Surgery,
The Emory Orthopaedic Center, Department of Orthopaedic Surgery, Emory University
School of Medicine, 59 Executive Park South, Suite 1000, Atlanta, GA 30329; e-mail:
claudius.d.jarrett@emory.edu.
0363-5023/14/3903-0003$36.00/0
http://dx.doi.org/10.1016/j.jhsa.2013.12.015
forces generated by a more vertical pattern in the
fracture line and the high tendency toward develop-
ment of a humpback deformity.2 Although initially
thought to be rare, recent data suggest that the oblique
scaphoid fracture pattern may be underappreciated.3
Furthermore, the dorsal sulcus fracture pattern ap-
pears to be the more common variant of the oblique
scaphoid fracture (Fig. 1).3
A current recommendation for stabilizing all sca-
phoid fractures, irrespective of the fracture pattern,
involves placing a screw down the central axis
of the scaphoid.4,5 This recommendation was based
on the improved biomechanical strength of a cen-
trally placed screw compared with an eccentric screw
when treating transverse waist fractures.4 Recent
studies suggest that it may be more appropriate to fix
oblique fractures by means of a screw that is directed

430 r Ó 2014 ASSH r Published by Elsevier, Inc. All rights reserved.

 SCREW TRAJECTORY IN UNSTABLE SCAPHOID FRACTURES 431

FIGURE 1: A Transverse waist scaphoid fracture (dotted line) on dorsal and volar views. B Dorsal sulcus pattern of scaphoid fracture (dotted
line) on dorsal and volar views. [Reproduced with permission and copyright © of the British Editorial Society of Bone and Joint Surgery from
Compson JP. The anatomy of acute scaphoid fractures: a three-dimensional analysis of patterns. J Bone Joint Surg Br. 1998;80(2):218e224.3]

perpendicular to the fracture line.6e10 One study used
a cadaveric biomechanical model to compare fixation
between perpendicular and central trajectories for the
oblique fracture.7 This study did not, however, apply
cyclic loading, nor did it simulate the more common
dorsal sulcus fracture pattern.3
The primary aim in this cadaveric study was to
determine whether a screw placed perpendicular to
the fracture line in an oblique scaphoid fracture could
achieve fixation comparable with that achieved by a
screw placed down the central axis. We surmised that
this perpendicular trajectory would also allow for a
shorter screw to be placed, thus providing the sur-
geon with an alternative to placing the longest screw
possible. A cadaveric biomechanical study was
designed to compare the effect of screw trajectory on
fixation strength for the oblique fracture pattern
following cyclic loading and load to failure. Second,
we sought to compare the average screw length and
mechanism of failure between the 2 groups.
MATERIALS AND METHODS
Cadaveric specimens
Eight pairs of scaphoids from the left and right wrists
of 8 fresh frozen cadavers were harvested. The
sample size was calculated assuming equal numbers
of specimens in both study samples, with a ¼ 0.05,

J Hand Surg Am. r Vol. 39, March 2014

432 SCREW TRAJECTORY IN UNSTABLE SCAPHOID FRACTURES
FIGURE 2: A Models depicting trajectory of screw placement in relation to dorsal sulcus/fracture line (red line). Left: longitudinal
screw. Right: perpendicular screw. R-S, radioscaphoid articulation. B Fluoroscopic view of perpendicular and central screw trajectories.
b ¼ 0.20 (to provide 80% power with 95% confi-
dence), and sD ¼ 0.40 (SD). This yielded a sample
size of 8, which was the sample size used in previous
studies in which at least 8 specimens provided
adequate power for statistical analysis.4,7,11,12 All soft
tissue attachments were dissected from the scaphoids.
None of the scaphoids showed any evidence of pa-
thology. One scaphoid in each pair was randomly
assigned to the central-axis screw group, and the
other was assigned to the perpendicular screw group.
This randomization resulted in half left and half right
scaphoids in each group. This protocol is consistent
with that used in previous studies.4,7,11
Fracture model
Oblique fracture patterns were then created in each
scaphoid using an identical technique specifically
devised for this study. In order to create a reproduc-
ible cut, the long axis of the scaphoid was identified.
A protractor was then used to indicate a line 45 to
the long axis and roughly in line with the dorsal
sulcus. The line for the osteotomy was drawn out
along the dorsal sulcus of the scaphoid to simulate the
oblique fracture pattern characterized in a recent
radiographic study.3 After the osteotomy line was
defined, the path for each screw was drilled prior to
making the cut. A thin-blade saw was then used to cut
along the marked line. After each osteotomy was
created, the fracture was anatomically reduced and
stabilized.
Fracture fixation
Fixation of the simulated fracture was achieved using
a standard screw from the Acutrac 2 Headless
Compression System (Acumed, Beaverton, OR).
Placement of the screw was done in accordance with
J Hand Surg Am.
the manufacturer’s technique guide. First, a guide
wire was placed either perpendicular to the fracture
line or down the longitudinal axis of the scaphoid.
A depth gauge was used to determine the appropriate
screw length. A cannulated drill was then inserted
over the guidewire and driven across the planned
fracture site. After the osteotomy, a screw of appro-
priate length was placed over the guidewire and the
fracture was firmly secured (Fig. 2). Screws were
inserted from proximal to distal using a volar starting
point. Appropriate screw length was confirmed using
visual inspection. Each specimen was potted in
cement with a Kirschner wire placed in the proximal
fragment to aid anchoring the specimen in the
cement. The proximal pole of the scaphoid was then
secured in cement and positioned so that the fracture
line was fully exposed. The specimen was oriented
with the long axis at 45 to horizontal in order to
simulate anatomical position of the scaphoid during
normal physiological loads (Fig. 3).4,13
Mechanical testing
The specimen was loaded from dorsal to volar with a
cantilever force, which has been shown to be the
primary physiological load encountered by the
scaphoid.4,14 A materials testing machine (Mini
Bionix, Eden Prairie, MN) was used to perform cy-
clic loading (Fig. 4). The specimens were loaded
from 80 N to 120 N at 1 Hz to simulate a subload to
failure at physiological load. These loading criteria
were based on prior studies and pilot testing.4,7 The
first 2 pairs were originally loaded with a maximum
cyclic force of 150 N. This was subsequently
changed to 120 N owing to early failure or dis-
placement, and the first 2 pairs were omitted from
analysis of cyclic loading but included in the analysis
r Vol. 39, March 2014
 SCREW TRAJECTORY IN UNSTABLE SCAPHOID FRACTURES
FIGURE 3: A protractor was used to measure a 45 angle to the
long axis of the scaphoid and in line with the dorsal sulcus in
order to create a reproducible line for osteotomy.
for load to failure. Testing was carried out until 2 mm
of fracture displacement occurred or 4,000 cycles was
reached. This displacement of 2 mm was used in a
previous study of similar design.4 The specimens that
reached the 4,000-cycle limit were then loaded to
failure while in the same position in the testing ma-
chine. Failure was defined as screw cutout or fracture
of the distal pole of the scaphoid. Mechanism of
failure for each specimen was recorded.
Data analysis
We compared the number of cycles loaded, the length
of the screw, and the force to failure between the 2
groups using the Wilcoxon signed-rank nonpara-
metric analysis. We then calculated the correlation
coefficient between screw length and load to failure
using both Pearson and Spearman (nonparametric)
analyses. Modes of failure were compared using the
Fisher exact test. Statistical significance was estab-
lished at P less than .05.
RESULTS
Biomechanical testing revealed no statistical differ-
ence in the number of cycles reached by the 2 groups:
3,510  1,199 cycles for the perpendicular screw
J Hand Surg Am.
433
FIGURE 4: Close-up view of load being applied through a
plunger to the distal pole of the scaphoid, which is potted with the
long axis at 45 .
group and 3,470  1,298 cycles for the central-
axis screw group (P > .05). One screw (16 mm
long) failed after 1,062 cycles at 120 N in the
perpendicular group, and 1 screw (20 mm long) in the
pilot longitudinal group failed after an initial load of
150 N was placed on the distal pole and before any
cycles were applied. Two fractures with longitudinally
placed screws underwent 2 mm of displacement
before cyclic testing was finished compared with 1
fracture with a perpendicularly placed screw.
There was no significant difference in load to
failure for the perpendicular screw group compared to
the central-axis screw group . There was a significant
difference in screw length between the 2 groups
(Table 1). Normalization of loads for screw length
showed no statistically significant difference between
the perpendicular group and the longitudinal group.
There was not a significant correlation between screw
length and load to failure with r ¼ 0.33 (P ¼ .23)
with the Pearson calculation and r ¼ 0.21 (P ¼ .46)
with the Spearman calculation.
The most common mode of failure of both groups
was fracture of the distal pole of the scaphoid. This
occurred in 6 of the 8 scaphoids in both groups. The
others failed by screw cutout. The distal pole was
the site at which the cantilever force was applied
and was away from the screw-bone interface. This
suggests that the bone itself failed before the screw
fixation did.
DISCUSSION
We sought to compare screws placed perpendicular
to the fracture line with those placed down the central
axis used to stabilize oblique scaphoid fractures. The
r Vol. 39, March 2014
434 SCREW TRAJECTORY IN UNSTABLE SCAPHOID FRACTURES

TABLE 1. Strength of Fixation and Screw Length

Trajectory Variable n Mean SD P

Central Screw length (mm) 8 22.8 2.1 .0089

Cycles 6 3,470 1,298 > .05
Load/screw length (N/mm) 8 12.9 4.7 .53
Load at failure 8 294 115 .92
Perpendicular Screw length (mm) 8 18.5 2.6
Cycles 6 3,510 1,199
Load/screw length (N/mm) 8 13.9 4.3
Load at failure 8 258 85

2 screw trajectories had similar responses to fatigue

testing, showing no statistical difference in number of
cycles loaded. Likewise, there was no significant
difference between the 2 groups when specimens
were loaded to failure. The perpendicularly placed
screws were significantly shorter than the axially
aligned screws. Load to failure in our study averaged
258 N for the perpendicular screw group and 294 N
for the central-axis screw group. These results are a
slight contrast to those obtained in cadaver and finite
element analysis studies.7e10 Luria et al7 studied load
to failure in an oblique cadaver model similar to ours
and found average loads of 137 N for screws placed
down the central axis and 148 N for screws placed
perpendicular (P > .05). The fracture pattern used in
the Luria et al study differed slightly in that it was a
short oblique pattern through the distal pole rather
than the dorsal sulcus pattern used in our study. This
FIGURE 5: Fluoroscopic extended oblique imaging depicts
short oblique pattern likely has different inherent proposed trajectory for perpendicular screw placement across
stability characteristics from the pattern we used, a dorsal sulcus oblique scaphoid fracture (red line). Technique is
which could explain the lower loads to failure performed with the wrist fully extended using a volar approach.
measured in that study. McCallister et al4 carried out
biomechanical experiments on the more stable
transverse waist repaired by means of a central screw. We believe that applying fixation perpendicular to an
They achieved load to failure of 513 N to 712 N.4 oblique fracture plane negates the need for a long,
The increased stability of the transverse waist fracture central screw.
pattern may explain the higher values for load to Central screw placement can be technically chal-
failure found in this study. lenging and is often limited by the patient’s body
There was a significant difference in screw lengths habitus and scaphoid anatomy and the surgical
between the 2 cohorts. Perpendicular placement approach.15e17 Using a distal volar approach often
allowed for a shorter screw with no drop-off in requires excising part of the trapezium, which risks
strength of fixation. Furthermore, when loads at cutout through the volar aspect of the scaphoid and
failure were normalized to screw length, loads at damage to the radiocarpal joint if the screw is too
failure remained similar. We did not find a significant long.15 Likewise, in a proximal dorsal approach,
relationship between increasing screw length and hyperflexion is required, which hinders orthogonal
strength of fixation. This is at variance with a previ- imaging and can lead to hardware breakage and
ous study done with central screw placement in screw prominence.16 The dorsal approach also puts
transverse waist fractures. That study found a linear numerous important structures at risk including the
correlation in fixation strength with screw length.5 posterior interosseous nerve, the extensor digitorum

J Hand Surg Am. r Vol. 39, March 2014

 SCREW TRAJECTORY IN UNSTABLE SCAPHOID FRACTURES
communis tendon to the index, and the extensor indicis
proprius tendon, which are on average 2.2 mm,
2.2 mm, and 3.1 mm, respectively, from the dorsal
guidewire.17 We propose directing a screw perpen-
dicular to the plane of an oblique fracture using a volar
and proximal approach in a fully extended wrist
(Fig. 5). Critical neurovascular structures and tendons
cross the volar surface of the wrist; therefore, further
investigation is required to evaluate the clinical utility
of this technique.
Additional benefits of a perpendicular screw may
extend beyond strength of fixation. Hart et al18
demonstrated that perpendicular fixation provides
greater potential surface area for healing in acute
oblique fractures. This represents an important adjunct
to healing in a bone with a tenuous vascular supply.
This study has several limitations. By nature of
studying cadaver bone stripped of soft tissue, it
is difficult to simulate the biomechanical in vivo
conditions. Also, only a limited number of speci-
mens could be tested secondary to the restrictive
access to cadavers. However, previous power ana-
lyses for similar cadaveric studies have shown that
a minimum of 8 specimens per screw group was
enough to yield statistically significant results with
an 80% confidence interval.4,7,11,12 We did not
standardize our models in regards to bone density
and age, although we did use matched pairs in order
to closely approximate the 2 cohorts. In previous
studies using paired cadaveric scaphoid models, no
difference was found between pairs in regards to
bone density.4,7,11
REFERENCES
1. Herbert TJ, Fisher WE. Management of the fractured scaphoid using
a new bone screw. J Bone Joint Surg Br. 1984;66(1):114e123.
2. Russe O. Fracture of the carpal navicular. Diagnosis, non-operative
treatment, and operative treatment. J Bone Joint Surg Am. 1960;42:
759e768.
J Hand Surg Am.
435
3. Compson JP. The anatomy of acute scaphoid fractures: a three-
dimensional analysis of patterns. J Bone Joint Surg Br. 1998;80(2):
218e224.
4. McCallister WV, Knight J, Kaliappan R, Trumble T. Central place-
ment of the screw in simulated fractures of the scaphoid waist. J Bone
Joint Surg Am. 2003;85(1):72e78.
5. Dodds SD, Panjabi MM, Slade JF. Screw fixation of scaphoid frac-
tures: a biomechanical assessment of screw length and screw
augmentation. J Hand Surg Am. 2006;31(3):405e413.
6. Moritomo H, Murase T, Oka K, Tanaka H, Yoshikawa H,
Sugamoto K. Relationship between the fracture location and the ki-
nematic pattern in scaphoid nonunion. J Hand Surg Am. 2007;33(9):
1459e1468.
7. Luria S, Lenart L, Lenart B, Peleg E, Kastelec M. Optimal fixation of
oblique scaphoid fractures: a cadaver model. J Hand Surg Am.
2012;37(7):1400e1404.
8. Luria S, Hoch S, Liebergall M, Mosheiff R, Peleg E. Optimal fixation
of acute scaphoid fractures: finite element analysis. J Hand Surg Am.
2010;35(8):1246e1250.
9. Ezquerro F, Jiménez S, Pérez A, Prado M, de Diego G, Simón A. The
influence of wire positioning upon the initial stability of scaphoid
fractures fixed using Kirschner wires: a finite element study. Med Eng
Phys. 2007;29(6):652e660.
10. Gokce V, Oflaz H, Dulgeroglu A, Bora A, Gunal I. Kirschner wire
fixation for scaphoid fractures: an experimental study in synthetic
bones. J Hand Surg Eur Vol. 2011;36(4):325e328.
11. Newport ML, Williams CD, Bradley WD. Mechanical strength of
scaphoid fixation. J Hand Surg Br. 1996;21(1):99e102.
12. Rankin G, Kuschner SH, Orlando C, McKellop H, Brien WW,
Sherman R. A biomechanical evaluation of a cannulated compressive
screw for use in fractures of the scaphoid. J Hand Surg Am.
1991;16(6):1002e1010.
13. Toby EB, Butler TE, McCormack TJ, Jayaraman G. A comparison of
fixation screws for the scaphoid during application of cyclical
bending loads. J Bone Joint Surg Am. 1997;79(8):1190e1197.
14. Smith DK, Cooney WP III, An KN, Linscheid RL, Chao EY. The
effects of simulated unstable scaphoid fractures on carpal motion.
J Hand Surg Am. 1989;14(2):283e291.
15. Levitz S, Ring D. Retrograde (volar) scaphoid insertion—a quanti-
tative computed tomographic analysis. J Hand Surg Am. 2005;30(3):
543e548.
16. Bushnell BD, McWilliams AD, Messer TM. Complications in dorsal
percutaneous cannulated screw fixation of nondisplaced scaphoid
waist fractures. J Hand Surg Am. 2007;32(6):827e833.
17. Adamany DC, Mikola EA, Fraser BJ. Percutaneous fixation of the
scaphoid through a dorsal approach: an anatomic study. J Hand Surg
Am. 2008;33(3):327e331.
18. Hart A, Mansuri A, Harvey E, Martineau P. Central versus eccentric
internal fixation of acute scaphoid fractures. J Hand Surg Am.
2013;38(1):66e71.
r Vol. 39, March 2014