Percutaneous pinning of ten fresh-frozen human cadaveric shoulders
was performed with the technique described by Jaberg et al.3 (Fig. 1). The specimens had previously
been disarticulated through the scapulothoracic joint. The proximal
part of the humerus was left intact as it was thought that creating
a fracture was unnecessary for the purposes of this study and that
it might compromise additional anatomic dissections. An identical pinning
technique was utilized for each specimen, and every effort was made
to insert all pins in a reproducible fashion.
With use of c-arm fluoroscopy, five 2.5-mm terminally threaded
AO pins were advanced into each specimen. Standardized fluoroscopic
views of the shoulder, including a true anteroposterior view of
the shoulder perpendicular to the scapular plane and a transscapular
lateral view of the humerus that was oriented 90° to the anteroposterior
view, were used during placement of the pins. Two lateral pins were
placed, from distal to proximal, through the lateral aspect of the humeral
shaft into the humeral head; two greater tuberosity pins were placed,
from proximal to distal, through the greater tuberosity, engaging
the medial aspect of the humeral neck; and one anterior pin was
placed, from distal to proximal, entering the anterior midline of
the humerus and aiming for the posterior aspect of the humeral head.
The anterior pin was placed perpendicular to the scapular plane
with the shoulder in neutral rotation, and the two lateral pins
and the two greater tuberosity pins were placed parallel to the
scapular plane. The lateral pins were advanced to within 5 mm of
the articular surface of the humeral head. The position of the tips
of the pins was evaluated with anteroposterior fluoroscopic views
with the shoulder in two, three, or four positions of rotation.
Every effort was made to avoid overpenetration of the humeral head and
shaft medially, as would be the case during an actual procedure.
In addition, satisfactory positioning of all pins was confirmed
with use of both anteroposterior and lateral fluoroscopic images,
as described above.
Each specimen was then dissected to determine the distance of
each pin from adjacent neurovascular structures, including the cephalic
vein, musculocutaneous nerve, axillary nerve (main trunk and anterior
branch), and posterior humeral circumflex artery. The distance of
the anterior pins from the tendon of the long head of the biceps
was also determined. Fine-point engineering calipers were used to
measure the shortest perpendicular distance between the pins and
the adjacent structures. Overlying muscles and soft tissues were
dissected when necessary to explore anatomic structures for measurements.
The means and ranges for these distances were then calculated and
recorded.
Other anatomic relationships that were measured included the length
of each humerus, the distance between the inferior aspect of the
articular cartilage of the humeral head and the axillary nerve,
the distance from the top of the humeral head to the cephalic vein
at the anterior midline of the humeral shaft, the distance along
the lateral aspect of the humeral shaft from the deltoid tuberosity
to the radial nerve distally, and the distance between the anterior
branch of the axillary nerve and the tip of the greater tuberosity
laterally. The humeral heads were inspected for any evidence of
pin penetration.
The distances of each pin from selected anatomic structures are
presented in Table I.
Lateral Pins
The distal lateral pins were located at a mean distance of 19 mm
(range, 10 to 25 mm) from the cephalic vein and a mean distance
of 17 mm (range, 13 to 25 mm) from the anterior branch of the axillary
nerve.
The proximal lateral pins were located at a mean distance of 21
mm (range, 14 to 27 mm) from the cephalic vein and a mean distance
of 3 mm (range, 1 to 10 mm) from the anterior branch of the axillary
nerve. In seven of the ten specimens, the proximal lateral pin was <2
mm from the anterior branch of the axillary nerve (Fig. 2). Four of the
twenty lateral pins were noted to penetrate the articular cartilage
of the humeral head.
Anterior Pins
The anterior pins were located at a mean distance of 11 mm (range,
0 to 19 mm) from the cephalic vein. One pin perforated the vein.
The anterior pins were also located at a mean distance of 2 mm (range,
0 to 7 mm) from the tendon of the long head of the biceps muscle.
In two specimens the pin perforated the tendon, and in one specimen
the pin passed between an anomalous bifurcation of the tendon. The
anterior pins were all located lateral to the musculocutaneous nerve,
at a mean distance of 17 mm (range, 9 to 35 mm).
Greater Tuberosity Pins
Distances were measured between the main trunk of the axillary
nerve and the posterior humeral circumflex artery and the tips of
the pins as they penetrated the medial aspect of the humeral neck
(Fig. 2).
These distances were measured with the shoulder in neutral rotation
as well as in internal and external rotation. In general, internal
rotation of the shoulder tightened the axillary nerve and the posterior
humeral circumflex artery and brought these structures closer to
the tips of the pins, whereas external rotation relaxed these structures
and tended to move them farther away.
With the shoulder in neutral rotation, the distal tuberosity
pins were located at a mean distance of 8 mm (range, 1 to 18 mm) from
the posterior humeral circumflex artery and 10 mm (range, 0 to 22
mm) from the axillary nerve. These distances decreased with internal
rotation, to a mean of 7 mm (range, 0 to 21 mm) from the posterior
humeral circumflex artery and 8 mm (range, 0 to 19 mm) from the
axillary nerve, and increased with external rotation, to a mean
of 14 mm (range, 4 to 25 mm) from the posterior humeral circumflex
artery and 15 mm (range, 4 to 26 mm) from the axillary nerve. In
one specimen, the tip of the distal pin touched the axillary nerve slightly
with neutral rotation and tented the nerve with internal rotation.
Likewise, in an additional specimen the tip of the distal pin was
found to be immediately adjacent to the posterior humeral circumflex
artery and to tent the artery with internal rotation.
With the shoulder in neutral rotation, the proximal tuberosity pins
were located at a mean distance of 7 mm (range, 1 to 10 mm) from
the posterior humeral circumflex artery and 6 mm (range, 2 to 12
mm) from the axillary nerve. These distances decreased with internal
rotation, to a mean of 4 mm (range, 0 to 11 mm) from the posterior
humeral circumflex artery and 3 mm (range, 0 to 10 mm) from the
axillary nerve, and increased with external rotation, to a mean
of 11 mm (range, 6 to 16 mm) from the posterior humeral circumflex
artery and 11 mm (range, 5 to 15 mm) from the axillary nerve. In
two specimens, the pin was noted to be immediately adjacent to the
artery and nerve and to tent these structures with internal rotation
of the shoulder.
Other Measurements and Observations
The mean distance of the anterior branch of the axillary nerve from
the superiormost aspect of the greater tuberosity, measured along
the lateral aspect of the humerus, was 50 mm (range, 43 to 57 mm).
The main trunk of the axillary nerve was located along the medial
aspect of the humeral neck at a mean distance of 10 mm (range, 7
to 14 mm) from the inferiormost extent of the humeral head cartilage.
In all specimens, the posterior humeral circumflex artery coursed
parallel and distal to the axillary nerve. Measured at the anterior
midline of the humeral shaft, the distance from the superiormost aspect
of the humeral head to the cephalic vein averaged 80 mm (range,
72 to 96 mm). The radial nerve was found in the spiral groove, distal
to the deltoid tuberosity by a mean of 31 mm (range, 28 to 35 mm)
as measured directly lateral on the humeral shaft. The mean length
of the ten humeri in this study was 329 mm (range, 275 to 350 mm).
The majority of proximal humeral fractures are inherently stable
and are readily treated with closed means1,2,5-8.
Internal fixation is indicated for displaced two-part fractures of
the surgical neck when acceptable reduction cannot be achieved or
maintained with closed methods, and in select three and four-part
fractures when the chances for fracture union and humeral head salvage
are reasonable1,2,5-8. While hemiarthroplasty
is generally preferred for head-splitting fractures and for four-part
fracture-dislocations, especially in older, low-demand patients9,10, internal fixation for complex
proximal humeral fractures is being increasingly recommended by
some authors11.
Many techniques have been described for fixation of unstable proximal
humeral fractures. The theoretical advantages of closed reduction
and percutaneous pinning include avoidance of devascularization
of fracture fragments, minimization of the risk of injury to the
blood supply of the humeral head, and reduced operative morbidity
by avoidance of an open procedure2,6,12,13.
Disadvantages of this technique include the potential for pin migration,
loss of reduction, and pin-site infection3,6,12,13.
Injury to important anatomic structures about the shoulder is also
a potential concern with this technique. The published technique
used in the current study may be associated with a risk of injury
to important anatomic structures about the shoulder girdle. Particularly
at risk are the main trunk of the axillary nerve and the posterior
humeral circumflex artery from the greater tuberosity pins, the
anterior branch of the axillary nerve from the proximal lateral
pin, and the cephalic vein, biceps tendon, and musculocutaneous
nerve from the anterior pin. Although the risk of injury to these
structures may be reduced by making a small skin incision and bluntly dissecting
to bone prior to pin placement, the proximity of some of these structures
to the pins in the current study suggests that certain alterations
in pin placement should be considered.
To avoid injury to the anterior branch of the axillary nerve,
the starting point for all lateral pins should be at or distal to
a point along the lateral aspect of the shaft equal to twice the distance
from the top of the humeral head to a line perpendicular to the
shaft at the inferiormost margin of the articular cartilage of the
humeral head (Fig. 3). On the basis of the dissections
performed in the current study, a safe zone was found between this
point and the deltoid tuberosity distally. The technical difficulty
associated with pin insertion increases as the starting point is
moved distally along the humeral shaft, and there is a risk of injury
to the radial nerve when pins are inserted distal to the deltoid tuberosity.
The lateral pins should be advanced no closer than 10 mm from the
articular surface of the humeral head. Despite our attempt to avoid
humeral head penetration in the current study by using two, three,
or four rotational fluoroscopic views, the articular cartilage was
penetrated slightly in two of the specimens. Therefore, multiple
fluoroscopic views of the glenohumeral joint should be used to ensure
that no pin has penetrated the joint surface.
The addition of an anterior pin has been shown to increase the rigidity
of fixation with percutaneous pinning4,14.
Which structures can be injured by an anterior pin depends on the
location of the pin relative to a plane that bisects the humerus
longitudinally and is perpendicular to the scapular plane. Pins
lateral to this plane can potentially injure the cephalic vein,
pins in this plane can potentially injure the long head of the biceps,
and pins medial to this plane can potentially injure the musculocutaneous
nerve. Thus, there is no truly safe zone anteriorly. While penetration
of the cephalic vein or the biceps tendon would not necessarily
lead to substantial functional loss, the surgeon must weigh the
potential biomechanical advantages of adding an anterior pin against the
substantial risk of vein or tendon perforation.
We recommend initial fixation of the humeral head fragment to
the shaft with lateral distal-to-proximal pins. Following this fixation,
the shoulder may be externally rotated during placement of greater
tuberosity pins to move the axillary nerve and the posterior humeral
circumflex artery farther away from the humeral neck. When the greater
tuberosity pins are inserted, great care must be exercised to avoid
overpenetration of the medial cortex, and the surgeon should aim
for a point 20 mm from the inferior extent of the humeral head to
reduce the risk of injury to the axillary nerve and the posterior
humeral circumflex artery (Fig. 3).
Potential weaknesses of this study bear mentioning. The cadaveric
specimens had been previously disarticulated through the scapulothoracic
joint. Theoretically, this may decrease tension in the brachial
plexus and vascular tree. However, because there was adequate tissue
proximal to the glenohumeral joint to tether the posterior cord
and the posterior humeral circumflex artery, we do not think that
the pertinent anatomic relationships measured in this study were compromised
substantially. The fact that we pinned intact rather than fractured
humeri may also be perceived as a weakness of the current study
because the actual clinical setting was not replicated. We believe
that it was important to leave the humeri intact for two reasons.
First, creating reproducible fracture patterns would have required
soft-tissue dissection and direct osteotomies of the proximal part
of the humerus. This may have distorted local anatomy and limited
the validity of subsequent anatomic measurements. Second, it is
generally assumed that displaced fractures will be reduced prior
to percutaneous pinning, thus restoring local anatomic relationships
to a nearly normal state. The presence of fracture hematoma or markedly
displaced fracture fragments may also alter the anatomic relationships
of the neurovascular structures studied in the current report. However,
it seems reasonable to conclude that the magnitude of such alteration
is likely to be small. Finally, while the modifications of the pin
positions suggested by this study are not dramatic, it should be emphasized
that these modifications were based only on anatomic considerations. In
vitro biomechanical studies and in vivo clinical
testing were not performed.
Percutaneous pinning of unstable fractures of the proximal part
of the humerus remains a viable technique. With slight modifications
in surgical technique, based on a knowledge of local anatomy, the
risk of iatrogenic injury to important anatomic structures about
the shoulder can be reduced.