Clavicle Fractures

A direct blow on the point of the shoulder is the most
common reported mechanism of injury that produces a midshaft fracture
of the clavicle.¹⁵,⁹⁵,¹³⁷,¹⁶⁰
This can occur in a number of ways, including being thrown from a
vehicle or bicycle, during a sports event, from the intrusion of
objects or vehicle structure during a motor vehicle accident, or
falling from a height. A recent prospective trial of more than 130
completely displaced midshaft fractures of the clavicle identified
motor vehicle/motorcycle accidents, bicycling accidents,
skiing/snowboarding falls or collisions, sports injuries, and falls as
the most commonly involved mechanisms.¹⁵
As the shoulder girdle is subjected to compression force directed from
laterally, the main strut maintaining position is the clavicle and its
articulations (Fig. 36-3). As the force exceeds
the capacity of this structure to withstand it, failure can occur in
one of three ways. The acromioclavicular (AC) articulation may fail,
the clavicle may break, or the sternoclavicular (SC) joint may
dislocate. SC injuries are rare and are typically associated with more
direct posterior blows against the medial clavicle (posterior
dislocations) or anterior blows to the distal shoulder girdle (levering
the proximal clavicle into an anterior dislocation).⁸¹,¹⁵⁸
Presumably, there are subtle nuances of the direction and magnitude of
applied forces and local anatomy that dictate whether the failure
occurs in the AC joint, or in the clavicle, and the magnitude of
displacement that occurs. Most (85%) clavicle fractures occur in the
midshaft of the bone where, as can be seen in a cross section, the bone
is narrowest and enveloping soft-tissue structures (which may help
dissipate injury force) are most scarce.²⁷,²⁸,²⁹,³⁰,¹³⁷,¹³⁸
It is typical to see a large abrasion or contusion on the posterior
aspect of the shoulder in patients with displaced midshaft clavicular
fractures, especially those who fall from bicycles, motorcycles, or
other vehicles: this force vector may also contribute to the location
of the fracture. The direction of the initial deforming force and both
gravitational and muscular forces on the clavicle are significant and
result in the typical deformity seen after fracture, with the distal
fragment being translated inferiorly, anteriorly, and medially
(shortened) and rotated anteriorly (Fig. 36-4).

Simple falls from a standing height are unlikely to
produce a displaced fracture in a healthy young person but can result
in injury in elderly, osteoporotic individuals: these fractures are
typically seen in the distal third of the clavicle. If the mechanism of
injury is trivial and does not seem commensurate with the fracture
depicted, then a careful investigation for a pathologic fracture should
be performed³⁰,¹⁵⁷ (Fig. 36-5).

Associated injuries are increasingly common in patients
with fractures of the clavicle, compared with the incidence reported in
older traditional studies.³⁶,⁴⁰,⁴¹,⁸⁹,¹⁶²,¹⁸³
There may be several reasons for this, including more liberal use of
improved diagnostic techniques (i.e., computed tomography [CT]
scanning), the greater speed and violence of many modern sports (e.g.,
motocross and snowboarding), and the improved survivorship of patients
with significant chest trauma who would have died before the
institution of comprehensive treatment of the trauma patient. In fact,
several studies from Level I trauma centers have examined the
characteristics of polytrauma patients with clavicle fractures and have
noted a high mortality rate (20% to 34%) from associated chest and head
trauma.⁸⁹,¹⁶²
Presumably, these series of critically injured patients contain
survivors who live to require treatment for the complications of their
clavicle fractures who may not have survived without modern trauma care.

Patients who have been the victims of high-energy
vehicular trauma are more likely to have associated injures to the
thoracic cage, including ipsilateral rib fractures, scapular and/or
glenoid fractures, proximal humeral fractures, and
hemothoraces/pneumothoraces²⁸,⁸⁹,¹⁶² (Fig. 36-6).
In addition to simply being good medicine, identification of these
injuries is important for multiple reasons. Patients may require urgent
treatment directed specifically at the associated injury (i.e., tube
thoracostomy for pneumothorax), their presence may influence the
treatment of the clavicle fracture (i.e., an associated displaced
glenoid neck fracture, the “floating shoulder” [see later]), or (as
objective information on this entity increases) they may give an
indication of the likelihood of a negative outcome for the clavicle
fracture (malunion, nonunion) that may have implications regarding
primary fixation (Fig. 36-7). The clavicle can also be injured from penetrating trauma including projectiles, blasts, and sword or machete blows (Fig. 36-8).
In this situation, diagnosing and treating underlying chest and/or
vascular injuries are critically important, and the clavicle can be
treated on its own merits.^* However, if a vascular repair
has been performed, clavicular fixation (if possible) provides an
optimally stable environment for healing.

The history should delineate a number of aspects to
optimize the patient’s care. In addition to the standard demographic
data, the details of the mechanism of injury are important. A clavicle
fracture caused by a simple low-energy fall is unlikely to be
associated with other fractures or intrathoracic injuries, whereas

a
fracture that occurs as a result of severe vehicular trauma or a fall
from a height should prompt a search for other injuries. In my
experience, clavicle fractures that result from falls while bicycling
often have associated multiple ipsilateral upper rib fractures. At a
Level I trauma center, McKee et al.⁸⁹
studied 105 polytrauma patients (multiple system injury and Injury
Severity Score greater than 16) with fractures of the clavicular shaft
and found a mortality rate of 32%, mainly as a result of associated
head and chest injuries. This high incidence of associated head and
chest injuries mandates careful clinical and radiographic
investigation. The physical mechanism of injury is important: while the
majority of fractures will result from a blow to the shoulder, failure
of the bone can also occur from a traction-type injury. This usually
occurs in an industrial or a dockyard

injury
in which the involved arm is forcefully pulled away from the body as it
is caught in machinery. It can also occur in vehicular trauma when the
arm is pinned against or strikes a fixed object as the torso continues
past it. This can lead to scapulothoracic dissociation, as the shoulder
girdle fails in tension at the SC joint, the clavicle, or the AC joint.
This is evident on the radiographs when a completely displaced,
distracted fracture site is seen (as opposed to the typical overlapping
fracture fragments) (Fig. 36-9).
The high incidence of neurologic and vascular traction injuries seen in
this setting mandates further investigation (i.e., angiography),
because they can be limb threatening.²⁹,³⁶,⁵⁴,⁹⁹,¹²²,¹⁸³

If the clavicular fracture has occurred with minimal trauma, one must be alert to the possibility of a pathologic fracture (Fig. 36-5).
Metabolic processes that weaken bone (i.e., renal disease,
hyperparathyroidism), benign or malignant tumors (i.e., myeloma,
metastases), or pre-existing lesions (i.e., congenital pseudarthrosis
of the clavicle) can result in pathologic fracture. In this setting,
nonoperative treatment of the clavicle fracture is recommended
initially, while intervention is directed toward diagnosis and
treatment of the underlying condition. Once the primary diagnosis has
been made and treatment initiated, the clavicle fracture is treated
based on its individual aspects. Also, repetitive or unusual loads may
induce a stress fracture of the clavicle, typically in bodybuilders or
weightlifters.¹¹⁹,¹⁴³,¹⁵²

In the past, when treatment of all clavicular shaft
fractures was consistently nonoperative, a detailed history of
lifestyle, occupation, and medical conditions was usually perfunctory
at best, since these factors did little to influence decision making.
However, there is increasing evidence that operative intervention is
superior in carefully selected cases of displaced clavicular shaft
fracture, such that additional information gleaned from the history
contributes to the risk/benefit analysis regarding possible surgery.
Compliant patients in the 16-to-60 age group, who have active
recreational lifestyles and/or physically demanding occupations
(especially those that require throwing, repetitive overhead work, or
recurrent lifting), are candidates for primary operative repair if they
are medically fit and have completely displaced fractures with good
bone quality.¹⁵,⁹⁶,¹²⁵,¹⁶⁷,¹⁸⁵
Factors associated with noncompliance and a high rate of fixation
failure, such as drug and alcohol abuse, untreated psychiatric
conditions, homelessness, or uncontrolled seizure disorders, are
contraindications for primary operative repair of clavicle fractures.¹⁰

When nonoperative treatment was chosen for the vast
majority of clavicle fractures, there was little emphasis placed on a
careful physical examination of the shoulder girdle. However, there are

a
number of findings that are important in surgical decision making.
There is usually swelling, bruising, and ecchymosis at the fracture
site, as well as deformity with displaced fractures. Visible deformity
of the shoulder girdle, best seen when the patient is standing, is an
important feature to recognize. The usual position seen with a
completely displaced midshaft fracture of the clavicle has been
described as shoulder “ptosis,” with a droopy, medially driven, and
shortened shoulder⁵⁷,⁶⁸,¹²²,¹³⁵ (Fig. 36-10).
In addition, the shoulder translates and rotates forward: this is a
deformity that can best be seen by viewing the patient from above.
Because of this malposition of the shoulder girdle, inspection of the
patient from behind may reveal a subtle prominence of the inferior
aspect of the scapula from scapular protraction as it moves with the
distal fragment. Shortening of the clavicle should be measured
clinically with a tape measure. A mark is made in the midline of the
suprasternal notch and another is made at the palpable ridge of the AC
joint: measuring this length gives the difference between the involved
and normal shoulder girdle.¹⁵⁵

A careful neurologic and vascular examination of the
involved limb is mandatory, especially if surgical intervention is
contemplated. If a deficit is not noted preoperatively, then it may be
incorrectly attributed to the surgery, which has prognostic,
medicolegal, and treatment implications.²⁷,²⁸,²⁹,³⁰

Surprisingly, given its subcutaneous nature and exposed
position, open fractures of the clavicle are relatively rare. Most open
fractures are associated with high-energy vehicular trauma, and
recognition is important for a number of reasons: the fracture itself
will require irrigation, debridement, and fixation, and there is a high
incidence of associated injuries. In the largest series in the
literature focusing on open clavicle fractures, Taitsman et al.
described 20 patients with this injury: 15 patients had pulmonary
injuries, 13 patients had head injuries, 8 patients had scapular
fractures, 11 patients had facial trauma, and there were a variety of
other injuries.¹⁶²

Simple anteroposterior (AP) radiographs are usually
sufficient to establish the diagnosis of a clavicle fracture. The
diagnosis may also be made from a single AP chest radiograph, which may
be the only available film in an urgent trauma setting. The chest
radiograph can also be used to evaluate the deformity of the involved
clavicle relative to the normal side and to look for associated
skeletal injuries such as rib, glenoid, and scapular fractures. A
measurement of length can be made on the chest radiograph comparing the
injured to the uninjured side: shortening of 2 cm or more represents a
relative indication for primary fixation. To best delineate a
clavicular fracture, as when one is determining whether operative
intervention is warranted, a radiograph should be taken in the upright
position (where gravity will demonstrate maximal deformity). Ideally,
the radiographic beam for the AP radiograph of the clavicle should be
angled 20 degrees superiorly to eliminate the overlap of the thoracic
cage and show the clavicle in profile.²⁷,²⁸,²⁹,³⁰,¹³⁴,¹⁷⁰
Also, if the torso is internally rotated a similar 20 degrees (rotating
internally when standing or by bumping up the opposite side while
supine), this places the scapula and shoulder girdle parallel to the
cassette for a true AP film. CT scanning of midshaft clavicular
fractures is rarely performed in the clinical setting, although this
imaging modality can demonstrate the complex three-dimensional
deformity that affects the shoulder girdle with these injuries,
including significant scapular angulation and protraction.⁵⁷
It is also useful for evaluating fractures of the medial third of the
clavicle and the remainder of the shoulder girdle, such as the glenoid
neck in cases of a “floating shoulder.”⁴²,¹³⁰,¹⁴²

Lateral clavicle fractures can be well visualized with
AP radiographs. Centering the radiograph on the acromioclavicular joint
and angling the beam in a cephalic tilt of approximately 15 degrees
(the Zanca view) helps delineate the fracture well, by removing the
overlap of the upper portion of the thoracic cage.²⁹,³⁰
To accurately delineate the degree of fracture displacement, these
radiographs should be taken with the patient standing and the arm
unsupported by slings, braces, or the uninjured arm. On occasion, it
may be useful to obtain a stress view to determine the integrity of the
coracoclavicular ligaments (as this can influence the choice of
fixation): a 5- to 10-pound weight is suspended from the wrist of the
affected arm and then radiographs

are
taken. CT scanning of lateral clavicle fractures is rarely required
clinically but can be useful in selected cases to determine
intra-articular extension or displacement.

Fractures of the medial clavicle, especially those
involving the SC joint, are notoriously difficult to accurately assess
with plain radiographs. CT scanning is the radiographic procedure of
choice when the anatomy of the fracture is unclear. This investigation
can help distinguish between a medial epiphyseal fracture (common in
individuals up to 25 years of age) and true SC dislocations²⁹,¹⁵⁰,¹⁶³,¹⁸¹ (Fig. 36-11).

A number of classification schemes have been proposed
for fractures of the clavicle. These have traditionally been based on
the position of the fracture, with the groups originally divided by
Allman into proximal (group I), middle (group II), and distal (group
III) third fractures. This general grouping has the advantage of
corresponding to the clinical approach to these fractures of most
orthopaedic surgeons.²⁹ Recognizing
that this basic scheme does not take into account factors that
influence treatment and outcome, such as fracture pattern,
displacement, comminution, and shortening, various authorities have
refined the classification to include other variables. Because of their
high rate of delayed and nonunion, Neer¹⁰¹
divided distal clavicle fractures into three subgroups, based on their
ligamentous attachments and degree of displacement (type II was
subsequently modified by Rockwood)²⁹:

Type I: Distal clavicle fracture with the coracoclavicular ligaments intact

Type II: Coracoclavicular ligaments detached from the medial fragment, with the trapezoidal ligament attached to the distal fragment

Type III: Distal clavicle fracture with extension into the AC joint

Ideally, a classification scheme should be reproducible
with a low rate of interobserver and intraobserver variability, should
help direct treatment, can be used to predict outcome, should be useful
in both the clinical and research realms, and should be simple enough
to be practically useful yet robust enough to include all fracture
patterns. While at the present time there is no classification scheme
that has been rigorously tested to meet all these objectives, modern
schemes based on prospective, comprehensive population-based studies
are available. Nordqvist et al.¹¹¹
examined more than 2035 fractures of the clavicle over a 10-year period
and essentially expanded on Allman’s original scheme by adding subtypes
based on fracture displacement, including a comminuted category for
midshaft fractures. In a similar population-based study in Edinburgh,
Robinson¹³⁷ evaluated more than 1000
consecutive fractures of the clavicle and developed a classification
scheme based on prognostic variables from the analysis of the data (Fig. 36-12).
It continues the traditional scheme of dividing the clavicle into
thirds and adds variables that are of proven diagnostic value
(intra-articular extension, displacement, and comminution). However, a
feature of this scheme is that it reverses the traditional numbering
scheme, describing medial fractures as type I, middle third fractures
as type II, and distal third fractures as type III. Because distal
third fractures are firmly entrenched in the orthopaedic lexicon as
“type II” fractures, this can lead to significant confusion. Despite
this drawback, the Robinson classification is based on an extensive
database that includes prospectively gathered, objective clinical data.
For this reason, it is the classification I prefer to use clinically as
it can help predict outcome and hence guide treatment, including the
decision to operate and fixation methods chosen. The AO/OTA Fracture and Dislocation Classification Compendium
was updated in 2007 to include recent developments including a unified
numbering scheme and measures to improve observer reliability (Fig. 36-13).
The clavicle is designated as segment 15 and is divided into the
standard medial metaphyseal, diaphyseal, and lateral metaphyseal
fractures.⁸⁵ An important difference
is that the metaphyseal fractures in this scheme are not one third of
the length of the bone but are shorter segments, according to the AO
“rule of squares.” For the all-important diaphysis, there are simple
(15-B1), wedge (15-B2), and complex (15-B3) subtypes.

Currently, there are two common surgical approaches
applicable to the fixation of clavicle fractures, each with its own
advantages and disadvantages, as follows:

There are numerous large series that describe relatively
good results following nonoperative treatment of clavicle fractures,
and it is my opinion that the majority of clavicle fractures can, and
should, be treated in this fashion.³,⁴⁰,⁴⁷,⁷²,¹⁰⁰,¹⁴⁴
However, there are serious deficiencies in these reports, including the
inclusion of children (who have an intrinsically good result and
remodeling potential), large numbers of patients lost to follow-up, and
radiographic and/or surgeon-based outcomes that are insensitive to
residual deficits. Recent evidence from prospective and randomized
clinical trials has suggested that there is a subset of individuals who
benefit from primary operative care¹⁵,⁶¹,⁶⁸,¹²⁵,¹³⁹,¹⁷² (Fig. 36-17).
Operative repair in this setting should be reserved for medically well,
physically active patients who stand to benefit the most from a rapid
restoration of normal anatomy and stable fixation. There are multiple
potential indications for primary operative fixation, outlined in Table 36-3.

The earliest reported attempt at closed reduction of a
displaced midshaft fracture of the clavicle was recorded in the “Edwin
Smith” papyrus dating from the 30th century B.C. Hippocrates described
the typical deformity resulting from this injury and emphasized the
importance of trying to correct it.¹
It is usually possible to obtain an improvement in position of the
fracture fragments by placing the patient supine, with a roll or
sandbag behind the shoulder blades to let the anterior displacement and
rotation of the distal fragment correct with gravity, followed by
superior translation and support of the affected arm. Unfortunately, it
is difficult or impossible to maintain the reduction achieved. For this
reason, over the millennia that followed the first description of
treatment of this fracture, there have been hundreds of descriptions of
different devices designed to maintain the reduction, including
splints, body jackets, casts, braces, slings, swathes, and wraps.¹,³,¹⁶,²⁷,⁸⁷
At the present time, there is no convincing evidence that any of these
devices reliably maintains the fracture reduction or improves clinical,
radiographic,

or
functional outcomes. For many years, the standard of care in North
America was the “figure-of-eight” bandage: Andersen and colleagues³
examined its utility in a prospective, randomized, controlled clinical
trial comparing it to a simple sling in 60 patients. They could
demonstrate no functional or radiographic difference between the two
groups, and in general the patients preferred the sling (2 of 27
dissatisfied with the sling compared with 9 of 34 dissatisfied with the
figure-of-eight bandage, P = 0.09). In a retrospective review of 140 patients treated nonoperatively, Stanley and Norris¹⁵⁹ did not find any difference between a standard sling and a figure-of-eight bandage, a finding confirmed by other authors.¹²³,¹²⁴
Also, I have seen several temporary lower trunk brachial plexus palsies
from figure-of-eight bandages that resulted from overtightening. For
this reason, in my practice, if nonoperative care is selected, a
simple, conventional sling with a padded neckpiece is applied, and no
attempt at reduction is made.

There are reports in the literature of various techniques of external fixation for clavicle fractures.³³,¹⁴⁸,¹⁶⁴
This method takes advantage of the intrinsic healing ability of the
clavicle and allows restoration of length and translation without the
scarring or morbidity of a surgical approach. Also, there is no
retained hardware at the conclusion of treatment. Schuind et al.¹⁴⁸
reported on a series of 20 patients treated with external fixation for
clavicular injuries, many of whom had local soft tissue compromise;
union occurred in all. Tomic et al.¹⁶⁴
described the treatment of 12 patients with nonunion of the clavicular
shaft by application of a modified Ilizarov device. Union was achieved
in 11 of 12 patients with an increase in the mean Constant-Murley
Shoulder Outcome Score from 30 preoperatively to 69 postoperatively. It
is clear that this technique is technically possible to perform and may
be useful in certain specific situations. Unfortunately, the practical
difficulties associated with the position and prominence of the
fixation pins, coupled with a lack of patient acceptance in the North
American population, has resulted in minimal use of this technique.

Intramedullary (IM) pinning of fractures of the shaft of
the clavicle has several advantages. These are similar to the benefits
seen with IM fixation of long bone fractures in other areas, although

this technique had not been as consistently successful in the clavicle as series in the femur or tibia have reported.⁸,²⁹,⁴⁵,⁴⁸,⁵⁶,⁶⁷,⁹⁴
Advantages include a smaller, more cosmetic skin incision, less soft
tissue stripping at the fracture site, decreased hardware prominence
following fixation, technically straightforward hardware removal, and a
possibly lower incidence of refracture or fracture at the end of the
implant. Recently, modifications to the technique have included a
radiographically guided completely “closed” technique.²⁴
Since, at the present time, there is no consistently reliable way to
“lock” an IM clavicle pin, complications include those common to all
unlocked IM devices, namely failure to control axial length and
rotation, especially with increasing fracture comminution and
decreasing intrinsic fracture stability. A biomechanical study of
clavicular osteotomies by Golish et al.⁵³
comparing 3.5-mm compression plates to 3.8-mm or 4.5-mm IM pins showed
that the plated constructs were superior in resisting displacement in a
number of different testing modes (maximal load, cyclical stress)
compared with both IM nail constructs.

The technique includes positioning the patient in a
semisitting position on a radiolucent table, with an image intensifier
on the ipsilateral side. By rotating the image 45 degrees caudal and
cephalad, orthogonal views of the clavicle can be obtained. A small
incision is then made over the posterolateral corner of the clavicle 2
to 3 cm medial to the AC joint (Fig. 36-18).
The posterior clavicle at this point is identified and the canal
breached with a drill consistent with the planned fixation device. A
reduction of the fracture is then performed, either through a small
open incision or, as experience increases, in a completely closed
fashion using a percutaneous reduction clamp on the medial fragment and
a “joystick” in the proximal fragment. Alternatively, the fixation
device can be inserted using a “retrograde” technique where it is
passed out from the fracture site through the lateral fragment. The
fracture is then reduced and the IM device is inserted into the medial
fragment under direct vision. It is important to accurately reduce
length and rotation, although the latter can be quite difficult if done
closed and no visual clues from the fracture configuration are
available. A small incision may be necessary to reduce vertically
oriented comminuted fragments and “tease” then back into alignment.
Following this, the canal is drilled to the appropriate size to accept
the planned IM device. Options include headed pins, partially threaded
pins or screws, cannulated screws, and smooth wires. Although some
series report favorable results with smooth wires, the North American
experience with smalldiameter smooth pin fixation includes breakage and
migration and is, in general, dismal.⁷⁸,⁸³,⁹⁷,¹⁰⁸
Smooth wires are contraindicated for fracture fixation about the
shoulder in general and for the clavicle in particular. It is important
not to distract the fracture site with the fixation device, which can
occur as the pin is inserted into the unyielding opposite cortex as the
S-shaped clavicle comes into contact with the end of the straight pin.
If this occurs, the pin must be withdrawn slightly or a shorter pin
used. The head of the pin or screw can be left prominent to facilitate
early removal through a small posterior incision or can be left flush
with the bone to decrease soft tissue irritation (Fig. 36-19).
Some authors advocate leaving the pin in a prominent position
subcutaneously for easy access in the clinic at the time of early (7 to
8 weeks postoperatively) hardware removal. This step depends on the
type of fixation device used and the philosophy of the treating
surgeon. The incisions are closed in a fashion similar to that used for
plate fixation, although they are typically smaller. If the surgeon is
confident with the stability of the repair, early motion is instituted
similar to that performed following plate fixation.

Infection had traditionally been one of the most feared
complications following fixation of displaced clavicular fractures, and
earlier series described an unacceptably high rate of deep infection.²,²⁷,⁵⁰,¹⁰⁰
However, significant improvements have been made in a number of areas
that are well recognized to decrease infection, including perioperative
antibiotics, selective operative timing with regard to soft tissue
conditions, better soft tissue handling, two-layer soft tissue closure,
and fixation that is superior biomechanically.^* In a recent
metaanalysis that examined operative series from 1975 to 2005,
Zlodwodzki et al. reported a superficial infection rate of 4.4%, and a
deep infection rate of only 2.2%; these figures are significantly
improved compared with earlier studies.¹⁸⁵
If infection does occur and it is superficial, then it is usually
possible to temporize with local wound care and systemic antibiotics
until fracture union has occurred. At this point, plate removal,
débridement, and thorough irrigation have a high success rate in
infection eradication.

Deep infection with unstable implanted hardware is a
more complex problem. If it appears that there is progressive bone
formation, then temporizing until union occurs followed by hardware
removal and debridement may be successful. If there is no obvious
progress toward union, then operative intervention is indicated.
Hardware removal followed by radical débridement of infected bone and
dead or devitalized tissue and subsequent irrigation is performed. At
this point there are several options. If the patient is healthy without
comorbidities (as is usually the case) and the infecting organism is a
sensitive one, then immediate reconstruction with plating, bone
grafting, and local antibiotics may be warranted. Alternatively,
especially with polymicrobial infections or resistant organisms (i.e.,
methicillin-resistant Staphylococcus aureus),
local antibiotic-impregnated polymethlymethylacrylate cement beads or
an antibiotic-impregnated bone substitute is implanted into any
residual dead space following debridement and systemic antibiotics are
administered until clinical and hematologic markers indicate the
infection has been eradicated. Delayed reconstruction can then be
performed. If there is a significant soft tissue deficiency, then the
assistance of a plastic surgeon who can perform soft tissue coverage,
typically with a rotational pectoralis major flap, is ideal.¹⁶¹,¹⁷⁴

Traditionally, the rate of nonunion of the clavicle has
been described as being less than 1% of all fractures. This was based
on two sentinel studies—one by Neer¹⁰⁰ in 1960 that described 3 nonunions in 2235 patients and one by Rowe¹⁴⁴
in 1968 in which only 4 of 566 patients developed nonunion after a
fracture of the clavicle. More recently, however, the nonunion rate
following closed treatment of completely displaced midshaft fractures
of the clavicle has been described as being exponentially higher, in
the 15% to 20% range.¹⁵,⁶²,¹³⁹,¹⁸⁵
The reason for this difference is unclear but probably includes more
complete follow-up in recent studies, the exclusion of children (with
their inherently good natural history), changing mechanisms of injury
(mountain biking, all-terrain vehicles, parachuting), and modern
patients’ intolerance of prolonged immobilization. In addition, several
prospective population-based studies have been helpful in elucidating
factors associated with the development of nonunion (Fig. 36-17). Robinson et al.¹³⁹
identified increasing age, female sex, fracture displacement, and
comminution as risk factors for nonunion in midshaft fractures. Lateral
third fractures had higher nonunion rates as patient age and fracture
displacement increased.¹³⁸ Nowak et al.¹¹³
prospectively followed 208 patients with radiographically verified
clavicle fractures and, 9 to 10 years postinjury, found that 96 (46%)
still had sequelae. They identified no bony contact between the
fracture fragments as the strongest predictor for sequelae. Nonunion
occurred in 15 patients (7%). Zlowodzki et al.¹⁸⁵
performed a metaanalysis of all series of displaced midshaft fractures
from 1975 to 2005 and identified 22 published manuscripts. They found
that, for the specific entity of completely displaced midshaft
fractures of the clavicle, the nonunion rate with nonoperative
treatment was 15.1%, while the nonunion rate following operative
treatment was 2.2%. This represents a relative risk reduction (for
nonunion) of 86% (95% confidence interval, 71% to 93%). This
meta-analysis, in addition to recent prospective studies examining
primary operative fixation of clavicle fractures, definitively
terminated the postulation that primary fixation was associated with a
higher, not lower, nonunion rate (Table 36-1). This observation was based on early operative studies with poor patient selection, inadequate fixation (Figs. 36-21 and 36-22),
and inferior soft tissue management. Undoubtedly, there are other
factors that contribute to the incidence of nonunion (i.e., associated
fractures, soft tissue interposition, rotation at the fracture site)
that have yet to be clarified.¹²⁹,¹⁵⁹,¹⁷¹,¹⁷³
Therefore, at the present time, factors associated with the development
of nonunion include complete fracture displacement (no contact between
the main proximal

and
distal fragments), shortening of greater than 2 cm, advanced age, more
severe trauma (in terms of both mechanism of injury and associated
fractures), and refracture. Primary operative fixation, however, is not
associated with a higher nonunion rate.

Nonunion is defined as the lack of radiographic healing at 6 months postinjury (Fig. 36-29). While a significant percentage of distal nonunions may be asymptomatic, especially in the elderly,¹⁰⁹ the majority of midshaft nonunions in young active individuals will be symptomatic enough to require treatment.^*

Traditionally, it was believed that malunion of the
clavicle (which was ubiquitous with displaced fractures) was of
radiographic interest only, and success in the clinical setting was
defined as fracture union. However, more recently, a number of
investigators have described a fairly consistent pattern of patient
symptomatology (with orthopaedic, neurologic, “functional,” and
cosmetic features) following malunion of displaced midshaft fractures
of the clavicle.⁷,²⁰,⁵⁷,⁶⁸,⁷⁶,⁹⁰,¹²²,¹³⁵
While all of the factors that contribute to the development of this
condition are unclear, it is typically diagnosed in young, active
patients with significant degrees of shortening at the malunion site (Fig. 36-31).
As could be reasonably anticipated, shortening of the shoulder girdle
(with the typical inferior displacement and anterior rotation of the
distal fragment) results in a variety of biomechanical and anatomic
abnormalities that translate directly into patient complaints.
Orthopaedically, shortening of the muscle-tendon units that traverse
the malunion site results in a sense of weakness and rapid
fatigability, with a loss of endurance strength. It has been previously
shown that there are significant, objective deficits in maximal
strength and endurance (especially of abduction) following the healing
of displaced midshaft fractures of the clavicle treated nonoperatively⁹²,¹⁵⁵ (Fig. 36-32).
Narrowing and displacement of the thoracic outlet (the inferior border
of which is the clavicle) result in numbness and paraesthesias, usually
in the C8-T1 nerve root distribution, exacerbated by provocative
overhead activities. Because of their deformity, patients complain of
the appearance of their shoulder and difficulty with backpacks, hiking
packs, military gear, and shoulder straps: this has been termed a
deficit in “functional cosmesis.” Patients with this condition also
complain of upper back pain and periscapular aching, especially with
repetitive activity. There is objective evidence that the displacement

of
the distal fragment (to which the scapula is attached) results in
malalignment of the scapulothoracic joint and a form of scapular
winging: this produces periscapular muscle spasm and fatigue pain.⁵⁷,¹³¹

It appears that the predominant risk factor for the
development of this condition is shortening at the malunion site. Hill
et al.⁶¹ found that shortening of 2
cm or more was associated with poor functional outcome and a high rate
of patient dissatisfaction. McKee et al.⁹⁰
described a series of 15 patients with a symptomatic clavicular
malunion who had a mean degree of shortening of 2.9 cm, and Bosch et al.⁹ described an “extension osteotomy” in four patients with shortening of 0.9 to 2.2 cm. In a retrospective study, Eskola et al.⁴⁷
reported on 83 patients with displaced fractures and found that
shortening of 1.2 cm or more was associated with increased pain at
final follow-up. However, this point remains controversial. In
retrospective reviews, Nordqvist et al.¹¹² (225 midshaft clavicle fractures) and Oroko et al.¹²⁰
(41 midshaft clavicle fractures) could not demonstrate a relationship
between shortening and a poor outcome. It is probable that length is
just one component of a complex three-dimensional deformity, which,
combined with the intrinsic variability of human response to skeletal
injury, explains why some individuals with malunion function well and
others determinedly seek operative correction. For patients with a
symptomatic malunion who have failed a course of physiotherapy for
muscle strengthening, the options are to accept the disability or have
a corrective osteotomy.

Despite the proximity of the brachial plexus and
subclavian vessels, neurovascular injury is surprisingly rare, given
the number of severely displaced clavicular shaft fractures seen in
practice.^* In general, neurovascular injuries associated
with clavicle fractures can be divided into three groups: acute
injuries, delayed injuries, and iatrogenic injuries.

True refracture after healing of a fracture of a
clavicle is surprisingly rare. It has been my experience that many
individuals who have claimed to have sustained multiple fractures of
the clavicle in fact, a nonunion following their initial fracture that
never healed completely. Recurrent episodes of trauma prompt medical
attention, and new radiographs are misinterpreted as showing a
“refracture.” The few cases that are reported describe a higher
nonunion rate following “refracture”: regardless of the exact etiology,
patients with this condition should be counseled about the high rate of
unsatisfactory outcome and that they may benefit from fixation.¹⁵,²⁹,³⁰

Given the increasing popularity of operative fixation of
displaced clavicle fractures and the patient population involved, it is
not surprising that fractures at the end of a plate used for fixation
of a prior clavicle fracture are being encountered. This typically
occurs from recurrent high-energy trauma. Large prospective series are
not available and recommendations are based on only a few cases. In
general, a fracture in the upper extremity that occurs at the end of a
stable implanted diaphyseal plate has a poor natural history and a high
chance of delayed union or nonunion. It is my experience that these
fractures, if displaced, generally require repeat fixation. Attempts
should be made to

fix the fracture and span the area of bone previously repaired (Fig. 36-36). If the fracture is minimally displaced, a trial of nonoperative care with the arm at rest in a sling is reasonable.

Winging of the scapula can take many different forms and
may have multiple etiologies, and has been anecdotally reported to
occur following the nonoperative treatment of displaced midshaft
fractures of the clavicle.¹³¹ Rasyid et al.¹³¹
reported two cases of winging of the scapula, one from a “neglected”
fracture of the clavicle with 2 cm of shortening. The typical clavicle
malunion or nonunion with deformity results in a shortened, protracted
shoulder. Since the scapula moves with the distal end of the clavicle
(through its only bony attachment at the AC joint), this can lead to
scapular malposition with a shortened, protracted shoulder (Fig. 36-37). This scapular malposition has recently been quantified in a series of patients with clavicle malunion by Hall et al.⁵⁷
using three-dimensional CT scanning in which mean clavicular shortening
of 16 mm resulted in 10 mm of lateral scapular displacement, with
anterior rotation of 8 degrees and 10 millimeters of elevation from the
posterior chest wall (Fig. 36-38). They postulated that this rotation and displacement result in the winging seen in their patients.⁵⁷
The negative mechanical effects of this shoulder position are well
documented, with a mean decrease in rotational strength ranging from
13% to 24% in one study.¹⁵⁶ This may
also explain the characteristic periscapular muscular fatigue pain that
these patients describe. Future research should elucidate the risk
factors for and potential treatment of this condition.

Recent studies have made it clear that there is a subset
of patients, especially those with shortened, displaced fractures, who
would benefit from primary operative repair of clavicular injuries.¹⁵,⁶²,¹³⁹
However, these early interventions are not without risk and consume
significant health care resources. Also, there are patients who appear
to have multiple prognostic factors for a poor outcome following a
clavicular fracture (i.e., displacement greater than 2 cm) and yet heal
promptly (albeit in a “displaced” position) with minimal symptomatology
and full function of the involved shoulder. While some of the
explanation for this is undoubtedly due to the inherent variability of
patient response to musculoskeletal injury, there may be other factors
that are not clearly defined or understood at present. For example,
while most studies use the magnitude of shortening when defining
fracture displacement, this alone is a relatively simplistic linear
measurement of what is typically a complex three-dimensional deformity.
Since most of the major muscle groups of the shoulder have a scapular
origin, it may be that the final position of the scapula relative to
the trunk and upper arm (which is difficult to measure; see “Scapular Winging”) may be the dominant factor in determining prognosis.⁵⁷ While the prognostic index published by Robinson et al.¹³⁹
is a dramatic advance in providing objective information and
facilitating our ability to predict outcome, there are still
significant improvements that can be made so that intervention can be
selected specifically for those patients for whom the risk/benefit
ratio of surgery is favorable. It is also clear that patient
noncompliance, especially when associated with substance abuse, is a
clear contraindication for surgery. Bostman et al.,¹⁰
in a study of 103 consecutive adults with acute, displaced midshaft
fractures of the clavicle, stated, “Patient noncompliance with the
postoperative regimen could be suspected to have been a major cause of
the failures.”

It remains unclear what the optimal method of fixation
is for displaced midshaft clavicle fractures in individuals who are
deemed appropriate for surgical fixation. There are proponents of both
IM fixation and plate fixation. The method of plate fixation is also
controversial, with some authorities recommending plate placement on
the superior surface of the clavicle, while others recommend the
anterior/inferior surface.¹⁰,¹⁵,²⁵,²⁷,⁷⁴,¹⁸⁴
There are no direct comparative studies between these various operative
techniques available at the present time, although it is hoped that
prospective studies that are currently under way comparing these
techniques will clarify the issue. While there are many theoretical
advantages to IM fixation, it appears that the results of this method
with currently available implants are more unpredictable than the
results reported for plate fixation. Biomechanically, they appear to be
inferior in resisting displacement compared with plate fixation.⁵⁵
Two clinical studies comparing IM fixation to nonoperative treatment
failed to show any advantage in the IM fixation group. Grassi et al.⁹⁶
described a high complication rate with IM fixation including eight
infections, three “refractures,” two delayed unions, and two nonunions
with hardware failure in 40 patients. Judd et al.,⁶⁸
in a randomized trial, failed to show an advantage of IM fixation over
nonoperative care in 57 patients, with nearly half the operative group
losing some degree of reduction. The metaanalysis by Zlowodzki et al.¹⁸⁵
did not reveal any significant differences between plate and IM
fixation, although this analysis was hampered by the lack of any direct
comparative studies. Conversely, Chuang et al.²⁴
described 100% union with no significant complications in a group of 34
patients with an acute midshaft fractures of the clavicle treated with
an IM cannulated screw. At the present time, IM fixation does not
appear to be as reliable in most series as plate fixation, and it
remains to be seen whether refinements in this technique and the
associated fixation devices can provide more consistent results.

It is probable that anteroinferior plating leads to less
plate irritation than placement of the plate on the superior surface of
the clavicle, at least with previously available plate systems. In the
one direct comparison (nonrandomized) between the two techniques, Lim
et al.⁸² reported a significantly better pain visual analogue scale for patients in the anteroinferior fixation group (P = 0.05). This finding awaits confirmation from further studies.

Conventional thinking has been that nonoperative
treatment is appropriate for most, if not all, fractures of the
clavicle, even severely displaced injuries, with the assumption that
the reconstructive repair of those that developed nonunion or
symptomatic malunion would produce results similar to that of primary
operative repair of the original fracture. Since these injuries are
nonarticular, and the reported “success” rate of reconstruction is
high, this approach seems to have inherent merit. However, there is
recent evidence that while operative reconstruction of malunion or
nonunion is a reliable procedure, with increasing refinement of outcome
measures and objective muscle strength testing, it is apparent that
there are residual deficits compared with what primary operative repair
can provide. Potter et al.¹²⁸
examined a cohort of 15 patients who had undergone late reconstruction
with plate fixation for clavicular nonunion or malunion (“delayed
group”) a mean of 63 months postinjury and compared them to a similar
cohort of 15 patients who had primary plate fixation of a clavicle
fracture a mean of 0.5 month after injury (“acute group”). The groups
were similar in age, sex, original fracture characteristics, and
mechanism of injury. They found that there were subtle but significant
differences between the two groups with regard to shoulder scores (a
Constant-Murley Shoulder Outcome Score of 89 in the delayed group and
95 in the acute group, P = 0.02), and the
delayed group demonstrated inferior endurance strength in the involved
shoulder. They concluded that while late reconstruction is a reliable
and reproducible procedure, there were subtle decreases in outcome
compared with acute fixation and recommended that this information be
used in decision making when counseling patients with displaced
midshaft fractures of the clavicle. Rosenberg et al.¹⁴²
reported similar results in a group of 13 patients who had late
reconstruction for clavicular malunion and nonunion. While osseous
healing occurred in all cases, there was a mean 20-point deficit in
Constant-Murley Shoulder Outcome Score (61 versus 81, P
= 0.01) in the affected shoulders. The authors thought that “lasting
functional impairment” was possible even with objective success. With
time, significant adaptive changes (muscle atrophy, soft tissue
contracture, bone loss) occur in the shoulder girdle of individuals
with clavicular malunion or nonunion that will compromise the outcome
of late reconstructive surgery to some degree compared with the results
of primary plate fixation. This is useful, objective information in
evaluating the risk/benefit ratio of early operative intervention.