Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49

Available online at

ScienceDirect
www.sciencedirect.com

Review article

Acute and chronic humeral shaft fractures in adults

L. Pidhorz ∗
Centre hospitalier du Mans, 194, avenue Rubillard, 72037 Le Mans cedex 9, France

a r t i c l e i n f o a b s t r a c t

Article history: Humeral shaft fractures account for up to 5% of all fractures. Many of these fractures are still being
Accepted 10 July 2014 treated conservatively using functional (Sarmiento) bracing or a hanging arm cast. Union is achieved in
10 weeks in more than 94% of cases. Angulation of less than 30◦ varus or valgus and less than 20◦ flessum
Keywords: or recurvatum can be tolerated by the patient from a functional and esthetic point of view. The ideal
Humeral shaft fracture candidate for this treatment is a patient with an isolated fracture. Plate and screw fixation of the fracture
Humeral anatomy results in union in 11 to 19 weeks. Reported complications include non-union (2.8–21%), secondary radial
Radial nerve palsy
nerve palsy (6.5–12%) and infection (0.8–2.4%). Anterograde or retrograde locked intramedullary nailing
requires knowledge of nailing techniques and regional anatomy to avoid the complications associated
with the technique. Union is obtained in 10–15 weeks. Reported complications consist of non-union
(2–17.4%), infection (0–4%) and secondary radial nerve palsy (2.7–5%). Hackethal bundle nailing is still
used for fracture fixation, despite an elevated complication rate (5–24% non-union and 6–29% pin migra-
tion) because of its low cost and simple instrumentation. Union is achieved in 8–9 weeks. Controversy
remains about the course to follow when the radial nerve is injured initially. If the fracture is open, sig-
nificantly displaced, associated with a vascular injury or requires surgical treatment, the nerve must be
explored. In other cases, the recommended approach varies greatly. Conservative treatment is inexpen-
sive and has a low complication rate. Humeral shaft fractures are increasingly being treated surgically,
at a greater cost and higher risk of complications.
© 2015 Elsevier Masson SAS. All rights reserved.

1. Introduction This review will focus on humeral shaft fracture studies, but will
exclude those involving diseased bone or that are periprosthetic in
Humeral shaft fractures account for about 5% of all fractures [1]. nature.
They almost exclusively occur in young people following a high-
energy trauma or older people following low-energy trauma. 2. Anatomy
A humeral shaft fracture is defined as one where the fracture line
is located between the insertions of the pectoralis major muscle Numerous muscles insert onto the humerus, which can explain
proximally and the brachialis muscle distally [2]. The AO defines the displacement of fracture fragments. Its medullary cavity is
a diaphyseal fracture of a long bone as one occurring between the funnel-shaped: the proximal portion has a larger diameter and rel-
two epiphyseal squares [3]. When these fractures are treated non- atively round shape; the distal portion is flatter and has a smaller
surgically, union is obtained in an average of 10 weeks, making the diameter. It has a very elongated S-shape on an oblique posterior
humerus a well-suited bone for conservative treatment. and medial plane, which corresponds to the humeral head retro-
The surgical indications (plate, nail, K-wire, external fixator) are version axis [4]. The axial torsion in the humeral shaft results in two
based on the surgeon’s school of training and presence of imme- smooth surfaces, a longer anteromedial one and a shorter antero-
diate complications (open fracture, radial nerve palsy). However lateral one (area where radial nerve passes through).
surgical fixation has many secondary and delayed complications Its anatomical relationship with three nerves is important to
associated with it (non-union, secondary radial nerve palsy, etc.) know when the fracture is being treated surgically [5]:
that require long treatment periods.
• radial nerve: it is in contact with the posterior side of the shaft;
this explains the high number of primary nerve palsy cases. It
∗ Correspondence at: 6, rue François-Coppee, 72000 Le Mans, France. passes in an oblique posterior groove from inside to outside
Tel.: +33 2 43 43 27 32/33 2 43 84 78 73; fax: +33 2 43 43 26 03. and superior to inferior over 6.5 cm. It crosses the lateral inter-
E-mail addresses: [email protected], [email protected] muscular septum at 16 cm from the lateral humeral epicondyle,

http://dx.doi.org/10.1016/j.otsr.2014.07.034
1877-0568/© 2015 Elsevier Masson SAS. All rights reserved.
S42 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49

which makes it vulnerable to displaced fractures in the mid- and distal third and is highly displaced. This is determined more on
dle third, particularly at the junction between the middle and the basis of the energy of the trauma than based on the radiographs,
lower thirds. Individual variations in the crossing point results in because the fracture can be realigned during transport or when the
a “danger area” located 10–15 cm from the lateral epicondyle [6]. radiographs are performed.
The exact position of the radial nerve was defined in a cadaver The neurological status cannot be determined in an unconscious
study [7]. It comes into contact with the posterior side of the patient.
humerus at 20.7 ± 1.2 cm from the medial epicondyle and leaves
it at 14.2 ± 0.6 cm from the lateral epicondyle; 4.2. Cutaneous trauma
• axillary nerve: it surrounds the posterior side of the surgical neck
from inside to outside, while following a horizontal arc 45 mm The skin overlying the fracture will be opened in 2–9% of cases
below the greater tuberosity of the humerus. It is flattened against following high-energy trauma.
the posterior side of the humerus;
• ulnar nerve: located behind the medial septum, it is near the pos- 4.3. Ulnar and median nerve palsy
teromedial edge of the distal part of the humerus, where it can
be damaged during the surgical approach or when using forceps. Ulnar and median nerve injuries are rare, but can be observed in
open fractures with significant muscle damage. One must be aware
The vascularization of the humeral shaft is heterogeneous. It is of the possibility of associated total or partial plexus injury (1.6–3%).
poor in the distal third, which can explain the greater number of Signs of radial nerve palsy must prompt the surgeon to look for axil-
non-unions at this level. lary nerve palsy; this combination is generally evidence of damage
The humerus is subjected mainly to rotational and distraction to the posterior fascicle of the brachial plexus. This is an important
forces; it is not subjected to compressive forces. The chosen fixation element to clarify before surgery and before regional anesthesia.
method must neutralize all of these rotational forces to achieve
union by first intention [8]. 4.4. Vascular injuries

3. Epidemiology [2,4,9–14] Vascular injuries are rare (0.5–3%). They are mainly due to
brachial artery rupture and require urgent care in collaboration
Humeral shaft fractures are the third most common type of with a vascular surgeon. The fracture must be stabilized before any
long bone fracture. Men are affected in more than half the cases blood vessels can be repaired.
(55–63%). The fracture occurs between 43 and 47 years of age, with
extremes of 15 and 97. However, there are two age clusters for these 4.5. Other associated injuries
fractures:
In high-energy trauma cases, the fracture may be associated
• 20–30 year-old males following high-energy trauma (motor vehi- with another musculoskeletal injury in the upper limb. The humeral
cle accident, fall from elevated high or sports injury); shaft fracture must be treated surgically first, so that any underly-
• 60–70 year-old women following low-energy trauma, such as a ing fractures can be treated without the risk of secondary damage
fall from her standing height. to the radial nerve.

The fracture line is located: 5. Additional assessments

• in the proximal third in 15–25% of cases and is often oblique; Two standard orthogonal A/P and lateral radiographs of the
• in the middle third in 49–64% of cases and is often transverse; entire humerus (including its ends) will be sufficient to identify the
• in the distal third in 11–35% of cases with increased incidence of fracture type, unless there are associated injuries. Traction X-rays
radial nerve injury. can be performed in the operating room under anesthesia if other
small fracture lines are suspected and need to be further evaluated
The fracture is simple in 56–63% of cases: spiral (18–29%), trans- before selecting a fixation method.
verse (21–32%) or oblique (11–15%). A third fragment is present in
26–34% of cases and the fracture is comminuted in 10% of cases. 6. Treatment

4. Clinical The treatment must match the fracture characteristics and com-
plications, along with the patient characteristics. In a 2003 French
A conscious trauma patient will present with the classical pic- multicenter study [4], surgical treatment had been used in 78% of
ture of an acute upper-limb injury. Immediate complications must fractures and non-surgical treatment in 22%. This distribution was
be identified [2,4,9,11,14]. affected by the recruitment of patients who often had multiple
fractures and by the nature of the participating hospital centers.
4.1. Radial nerve palsy
6.1. Conservative treatment
Radial nerve palsy is present initially in 10–20% of fracture
patients and typically manifests itself as paresthesia/paralysis of This consists of immobilization of a non-displaced or reduced
wrist dorsiflexion, finger extension at the metacarpophalangeal fracture with or without anesthesia. This is the preferred treatment
joints, thumb extension and abduction and hypo-/anesthesia of the of many surgical teams, despite the progress made with surgical fix-
dorsal side of the first inter-digital corner. ation methods. As noted by Chauveaux [4], this treatment method
Radial nerve involvement must be pointed out to the patient is constrained by the need for bracing and the current patient’s
and/or family, and recorded in the observations. It plays an impor- expectations.
tant role in the treatment choice and follow-up. The fracture is All types of humerus fractures can be treated, even cases with
typically located in the middle third or at the junction of the middle radial nerve palsy. However, this treatment strategy is challenging
 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49 S43

Fig. 1. Conservative treatment using functional (Sarmiento) bracing of a 3-fragment humerus shaft fracture without radial deficiency. (a) At admission. (b) With the functional
brace at day 14. (c) At the third month post-injury.

for the surgeon and patient, as weekly follow-up visits are required its alignment [16], and that micro-movements at the fracture site
early on. It also requires that the patient adhere to the instructions. contribute to bone union (Fig. 1).
Some surgeons believe these methods are outdated. The treatment initially consists of immobilization with an arm-
to-chest bandage to help align the fragments, without trying to
anatomically reduce the fracture. This initial immobilization period
6.1.1. Hanging arm cast
lasts 9 to 20 days [11,17]. In the next step, a plastic or resin device
This is a “classic” treatment method as evidenced from the De
is custom made for the patient’s arm by an orthotist. Velcro straps
Mourgues [10] and Babin [15] publications, but it was still used in
are used to regularly tighten the brace, ideally during daily visits to
50% of patients treated conservatively in a 2003 French multicenter
the clinic, but in reality every week. An arm-to-chest bandage can
study [4].
be added for 2–3 weeks to help reassure the patient. Patients must
The fracture is reduced because of the traction induced by the
make regular muscle contractions once the sleeve is applied and
weight of the long-arm cast (1–1.5 kg). Patients must be able-
must move their shoulder and elbow later on. This treatment can
bodied and well-informed, so as to let the casted arm hang, carry
be started in the emergency room on an outpatient basis, without
out pendulum movements of the shoulder and let the cast rest on
the need for hospitalization. The sleeve must be worn until union
their chest at night. The cast is worn for at least 6 weeks. Union
is achieved after an average of 11 weeks (range: 5–22) [11,17,18].
is obtained after an average of 52 days (7.5 weeks). The non-union
Sarmiento [11] reported a 43.7% rate of more than 5◦ varus angu-
rate is between 2 and 5% [4,10,15]. There is a risk of shoulder and
lation (with 12% having more than 16◦ ) and 1.5% rate of more than
particularly elbow stiffness developing, along with neck pain due
5◦ valgus angulation (all less than 16◦ ) on A/P radiographs. On lat-
to the weight of the cast.
eral radiographs, 16.1% of cases had anterior angulation greater
than 5◦ (4.8% > 16◦ ) and 13.9% with posterior angulation greater
6.1.2. Functional (Sarmiento) bracing than 5◦ .
This method is based on the fact that muscle pressure induced Zagorski [18] reported a similar portion of more than 5◦ varus
by the circular device gradually reduces the fracture and preserves or valgus angulation in his study. He also reported an average
S44 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49

shortening of 4 mm (range: 0–15), without functional conse-

quences.
Because of the shoulder’s mobility, malunion can be tolerated
since there are no functional or esthetic consequences. Thus, a varus
or varus angulation of less than 30◦ , or flessum or recurvatum of
less than 20◦ can be tolerated. The shoulder function is normal
in 70–88% of cases and the elbow function is normal in 76–88%.
Rotational malunion and its potential repercussions on internal and
external rotation movements of the shoulder have not been evalu-
ated. Union is obtained in 94–98% of fractures, depending on if they
are open or closed [4,11,17,18].

6.1.3. Other non-surgical treatment options

Other non-surgical treatment options are:

• the shoulder spica cast [15] dates back many years. It has very
little use now because of its bulk and secondary stiffness [12].
• transolecranon traction at the apex can be used as a temporary
treatment [15].
• arm-to-chest bandage (“Dujarrier” or “Mayo-clinic bandage”)
requires that the device be changed regularly and is not well-
tolerated by patients because of the two-month treatment
duration; the resulting reduction and bracing are not satisfactory
but this method might be required under certain circumstances.

Fig. 2. Treatment with locking plate of a highly displaced fracture of the middle
The best indication for conservative treatment is a patient with
third of the humeral shaft. (a) At admission. (b) At union.
an isolated humeral shaft fracture. It remains the gold standard
treatment in this population. Radial nerve palsy is not a contra-
indication. Union is nearly guaranteed; angular deviations are well- be noted in the operative report. In a variation of this approach
tolerated functionally, without repercussion on shoulder and elbow when the two median quarters are fractured [19], the plate can
mobility and with few complications. However, some have pointed be applied on the medial aspect of the bone through a curved
out that the immobilization period is too long and uncomfortable anterolateral incision passing between the biceps brachialis and
for the patient. the anterior brachial muscles. The elbow is flexed and the arm
externally rotated. The location of the radial nerve does not need
6.2. Surgical treatment to be known when using this approach. No postoperative com-
plications occurred with this approach. Union is obtained after
The goal of surgical treatment is to obtain anatomical reduc- an average of 80 days and the outcomes are satisfactory in 89% of
tion, while providing stability that allows for early mobilization cases.
of adjacent joints. It has its place in multifracture or polytrauma • posterior approach: the presence of the radial nerve makes this
patients, open fractures, failed conservative treatment and obese approach challenging with the patient prone. In a cadaver study
patients. Its indications have expanded over the last 20 years due of three types of posterior approaches [7], the preferred approach
to pressure from patients who want a treatment that allows them consisted of medially spreading of the medial and lateral heads of
to quickly return to their activities or who refuse to put up with the the triceps brachii by sectioning the lateral intermuscular septum
inconveniences of conservative treatment, while accepting the risk over 3 cm, which provided a 26.4 ± 0.4 cm long exposure of the
associated with surgery. humerus.
• minimally invasive approaches: these can be either lateral or
6.2.1. Plate fixation anterior (more reliable) and have been used since minimally inva-
This is a reliable, well-established technique used in 20.7% of sive techniques were introduced. Detailed knowledge of radial
cases in the 2003 French multicenter study [4] and 30% of cases in nerve anatomy is indispensable for these approaches.
the 1997 French multicenter study [12] making it the second most • anterior transposition of the radial nerve through the fracture has
commonly used method. It requires a very rigorous technique to been proposed [20]. This option is useful in middle third fractures
minimize the associated complications (Fig. 2). of the humeral shaft, as it increases the length by 11 mm, mak-
ing easier to release and move the nerve vertically and then to
6.2.1.1. Surgical approaches. The choice of approach is dictated by perform the fixation.
the fracture location, the preoperative status of the radial nerve
and the surgeon’s experience. The patient is placed supine with the 6.2.1.2. Standard plates. These plates are thick, fairly narrow
upper limb on an arm board, without a tourniquet: (4.5 mm), use 3.5 mm screws and provide similar compression of
the fracture site as dynamic compression plates. The construct must
• medial approach: it allows the humeral artery to be checked in incorporate six to eight cortices on either side of the fracture.
cases of vascular injury; it is the logical approach for plating About 25% of patients treated with plates in two French mul-
because it is far away from the radial nerve [4]. ticenter studies [4,12] had preoperative radial nerve palsy, which
• anterolateral approach: the most commonly used approach [12]; led to the decision to use a plate. The complications were non-
it allows the radial nerve to be located. The plate is applied to the union (8% and 21%, respectively), secondary radial nerve palsy (8%
middle part of the anterolateral side of the humerus shaft. Partic- and 12%), fixation failure (1.5% and 14%) and infection (0.8% and
ular attention must be paid to the nerve, which is dissected over 2%). These rates were substantially different to those reported in
the entire length of the incision. Its position relative to plate must single-center studies, for example the Paris et al. study [14] with
 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49 S45

Fig. 3. Anterograde intramedullary locked nail. (a) Seidel nail inserted into closed fracture with incision made to release the radial nerve (immediate palsy). (b) Static locked
nail inserted into closed fracture with freehand anteroposterior distal locking.

5% non-union, 5% secondary radial nerve palsy, 6% fixation failure to avoid any impingement between its proximal end and the
and 1% infection. His analysis of published studies revealed an aver- acromion.
age non-union rate of 2.8%, a 6.5% secondary radial nerve palsy rate Proximal locking is carried out with frontal or sagittal screws. It
and a 2.4% infection rate. must avoid the axillary nerve. The frontal screws through the head
Union is obtained after 11 to 19 weeks. The functional outcomes must not be too long to avoid entering the joint. Two pitfalls must
were good or very good in 96% of cases in one study [12] and 86.3% in be avoided when using a sagittal screw: crossing the long head of
the other [14]. In one of the French multicenter studies [4], the func- biceps tendon in front and injuring the axillary nerve in back while
tional elbow range of motion was normal and shoulder abduction drilling the screw hole or inserting an overly long screw. Trans-
was normal or subnormal in more than 90% of cases. verse distal locking can damage the radial nerve, thus the screw
must be placed in the anteroposterior direction. A short approach
is recommended to avoid these complications.
6.2.1.3. Locking compression plates. It has been said that locking
Of the 19% of fractures that were treated by IM nailing in a French
compression plates (LCP) provide no biomechanical advantage over
multicenter study [4], there was a 17.4% non-union rate, 4% infec-
standard plates, except probably in osteoporotic subjects [1]. No
tion rate and 2.7% postoperative palsy rate. These findings were
studies have specifically looked at the use of LCPs in humeral shaft
similar to the ones reported by Asencio et al. with locked nails [13]:
fractures.
10% non-union, 2.6% postoperative radial palsy, no infection. Union
was obtained after an average of 11 weeks (range: 4–40).
6.2.1.4. Plate removal. Removal of a plate must take into account After comparing standard plates with locked nails in a meta-
the risk of postoperative palsy, which is not insignificant. analysis, Kurup [24] concluded that despite an increased risk of
shoulder pain and stiffness, and the need to remove the hardware,
there was not enough evidence in favor of either type of fixation,
6.2.2. Intramedullary nailing
even in terms of function.
Intramedullary (IM) nailing was developed by Kuntscher in the
1940s and widely disseminated by Seidel [21]. Distal locking of Sei-
del’s nail made use of expandable fins. Long nails with distal screw 6.2.2.2. Retrograde nails [25,26]. These nails are introduced
locking were introduced in the early 2000s. These prevent tele- through the mid-line posterior triceps splitting approach. Fracture
scoping of the fracture fragments and rotational malunion. These of the distal end of the humerus during insertion or extraction is a
have replaced Seidel’s nail [22], which required additional immo- risk that is specifically associated with this type of nail.
bilization because of the precarious distal locking. This technique When using a Marchetti nail, Butin et al. [25] reported no cases
requires a good understanding of IM nailing and rigorous methods of infection or postoperative palsy, however there was a 5% rate of
to avoid complications [21,23]. posterior cortex cracks, 5% rate of supracondylar fractures upon nail
removal and 5% rate of non-union. Union was obtained in 10 weeks
(range: 6–16).
6.2.2.1. Anterograde nails. These nails are introduced through the
Apard et al. [26] used a static locking retrograde nail. They
greater tuberosity of the humerus under fluoroscopy control with
reported a 5% rate of postoperative palsy (regressive in 6 months),
the patient in the beach chair position and the arm in retropulsion
3.5% rate of supracondylar fracture and 2% rate of non-union. Union
(Fig. 3). The fluoroscopy unit must be placed in the correct position
was obtained in 15 weeks (range: 6–28). The overall functional out-
before the procedure.
comes were excellent in 86.6% of cases.
The insertion point is key; it will be in the axis of the shaft at
the base of the muscle–tendon junction of the rotator cuff or the
cartilage–upper anatomical neck junction, with the incision made 6.2.3. Other fixation methods
along the fibers of the rotator cuff. 6.2.3.1. Hackethal bundle nailing [4,12,27]. This retrograde bundle
The main criticism of anterograde nails lies in the approach nailing technique is performed percutaneously through the supra-
through the rotator cuff. The nail must be driven down sufficiently olecranon approach (Fig. 4). The goal is to fill the shaft with K-wires
S46 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49

Fig. 4. Treatment by Hackethal bundle nailing of a humerus shaft fracture. (a) At admission. (b) Immediately postoperative. (c) Non-union with K-wire failure; the patient
did not want an additional procedure (radiographs at 5 years).

to stabilize the fracture, separate the K-wires in the humeral head with or without vascular complications (1–2.2% of cases in pub-
to increase the construct stability and lock them distally to prevent lished series [4,12]) (Fig. 5). For some surgical teams, this is a
downward migration. Several variations of this technique have temporary treatment in the context of damage control surgery for
been described [28]. polytrauma patients.
The main complications are non-union (24% for Lefevre [4], 8% The bicortical pins are inserted on the lateral side of the humerus
for Nieto [12], 7.3% for De La Caffiniere [28] and 5% for Gayet [27]) under fluoroscopy control while keep the nerve locations in mind.
and upward or downward migration of the K-wires (6% [27] to 29% The proximal pins must be inserted distally to the axillary nerve;
[28]), which requires early removal. Non-union is due to a technical the distal pins must avoid the “danger area” associated with the
error; an inter-fragment gap of more than 3 mm was found to have radial nerve. Making a small incision will help to prevent nerve
a statistically significant negative effect on the outcome [4]. K-wire damage in cases where the pins are inserted into at-risk areas.
migration is secondary to poor filling of the shaft and/or inadequate Single-rod type fixators are preferred over Hoffmann-type ones, so
impaction of the fracture [4]. Conversely, there were no fractures as to obtain the most stable construct possible. The body of the fix-
at the insertion point in any of these studies. The shoulder func- ator must be located as close as possible to the axis of the humeral
tion was normal in 87% of fracture cases and the elbow function shaft. The pin openings must be cared for very carefully.
was normal in 63% [4]. Gayet et al. [27] reported that normal arm The device is worn for 14 weeks, which corresponds to the aver-
function had been achieved in 94.4% of cases; Nieto et al. [12] had age time to union. The elevated non-union rate with this device is
similar findings (92%). Union was obtained in 8–9 weeks [4,27,28]. likely due to the type of fractures being treated (i.e. the most com-
De La Caffinière et al. [28] preferred using this technique when plex ones). No postoperative complications related to the radial
the middle third of the humerus shaft is fractured, as these cases nerve were reported in a French multicenter study [4].
allow for more efficient filling of the shaft. Lefèvre [4] advised This is a technique that must be mastered; the indications are
against using this method with distal, comminuted or bifocal frac- rare and the insertion technique is very demanding because of the
tures, and fractures in polytrauma patients. risk of nerve damage.
The number of complications must be weighed against the
advantages of this type of treatment (low cost, no specific instru-
7. Radial nerve palsy
mentation needed, fast to carry out). This can be considered an
enhanced conservative treatment method, requiring an additional
This complication continues to be controversial. Should the
arm-to-chest bandage until signs of bone union are evident.
nerve be systematically examined and should we abandon conser-
vative treatment, given that this treatment leads to spontaneous
6.2.3.2. External fixator [2,4]. The indications for use of an external recovery in 85–100% of cases [1,4,17]? Should the nerve be sys-
fixator are rare and limited to open highly-contaminated fractures, tematically examined when performing surgical fracture fixation?
 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49 S47

Fig. 5. Treatment with single-plane external fixator of an open Gustilo Type 3A fracture with immediate radial nerve palsy. (a) At admission. (b) Single-plane external fixator.
(c) At the 15th day. (d) After removal of fixator at the 3rd month.

The possibility of radial nerve palsy must be considered at the ◦ if the nerve is only contused and still responds to stimulation,
various treatment stages; it must be reconsidered if secondary it is left as is and the fracture fixation material is placed away
palsy develops. It is most often due to neuropraxia and rarely to from the nerve,
nerve division. ◦ if it is empty and cannot be stimulated or obviously divided,
The recommendations of the 2003 SOFCOT symposium [4] were direct suture, resection-grafting or identification of the ends to
the following: facilitate secondary repair is indicated, according to the sur-
geon’s competencies;
• other cases: monitor the clinical progression and carry out an
• cases with immediate radial nerve palsy: evaluate the condition electromyography (EMG) exam around day 45.
of the nerve in cases of skin opening or vascular damage, which
provides information as to its macroscopic condition; if the nerve Alnot et al. [29] believe this nerve must be evaluated system-
is continuous, palpate and stimulate the nerve: atically when the fracture is highly displaced. In non-displaced or
S48 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49

minimally displaced fractures, nerve evaluation is only carried out if of inflammation possible, thereby reducing peripheral fibrosis
recovery does not occur in 2–6 months. In “intermediate” cases, the [29,30,8,31]. The two procedures must be spaced at least 6–8
procedure is appropriate in cases of polytrauma and high-energy weeks apart. Union is obtained in 80% of cases in an average of
trauma. The fact that the results are better following primary suture 16 weeks [8] to 33 weeks [12].
repair than secondary repair with grafting supports this approach.
Postoperative radial palsy has a good prognosis with fast
8.2. Infection
recovery, except in cases of intra-operative injury. Even a small
intra-operative trauma will further add to the initial nerve injury
This is a fact of life for any surgical treatment, but especially
[29]. The radial nerve palsy rate after various treatments reported
plate fixation. The infection rate reported in published studies was
in published studies was on average 1.4% for conservative treat-
on average 4% for plating, 0.8% for bundle nailing, 1.6% for locked
ment, 6.5% for plating, 1.7% for bundle nailing, 1.9% for locked IM
IM nails and 4% for external fixation [14].
nails and 1.4% for external fixation [14]. It is recommended that the
A microbiological diagnosis is essential before antibiotic treat-
nerve be identified before performing IM nailing [13,22]. In cases
ment is initiated. The treatment consists of removing the current
of plate fixation, the position of the nerve relative to the plate must
fixation device, excising the infected tissues and then applying an
be accurately defined.
external fixator.
In cases of radial nerve palsy without nerve evaluation or of sec-
In cases of septic non-union, removal of the internal fixation
ondary palsy, clinical and EMG monitoring every 6 weeks is needed
device in combination with excision of the infected tissue is carried
to follow the recovery and determine if nerve exploration is needed.
out before an external fixator is applied. Antibiotic therapy will be
EMG is pointless in the early days after the fracture.
adapted to the micro-organisms identified in consultation with an
infectious disease specialist. The treatment duration is long and the
8. Late complications outcome uncertain.

8.1. Non-union
9. Conclusion
Non-union is defined as the fracture not having healed in
Although humeral shaft fractures are often treated conserva-
6 months [2,28] (Fig. 4c). The non-union rate varies between 3 and
tively, the progress made with internal fixation devices and the
20%, depending on the treatment. The non-union rate reported in
pressure from patients have led to increased use of surgical treat-
published studies was on average 4.4% for conservative treatment,
ment. Bone union is faster with conservative treatment (9 weeks)
2.8% for plating, 6.3% for bundle nailing, 5.9% for locked IM nails
than plate fixation (11–19 weeks).
and 3.5% for external fixation [14].
Conservative treatment is the least costly treatment option
It occurs almost exclusively following transverse fractures of the
because it is performed on an outpatient basis (hospitalization not
middle third of the shaft [28], and the risk increases due to technical
required); the complication rate is low (especially for non-union)
errors:
and functional recovery is fast.
Internal fixation leads to more anatomical reduction, but the
• persistent inter-fragment gap of 6.3 ± 4.5 mm following func-
trade-offs are iatrogenic complications such as infection, fixation
tional bracing [30]; failure, secondary radial nerve palsy and non-union. All the internal
• insufficiently stiff plate construct with insufficient number of cor-
fixation methods are burdened by complications and their results
tical fixation points; are comparable. No matter which method is chosen, meticulous
• failure to lock nail or use of overly small nail diameter;
technique is essential to reduce the complication rate.
• insufficient filling of shaft by K-wires.

Disclosure of interest
This condition requires surgical treatment to apply stable fixa-
tion, with or without a graft.
The authors declare that they have no conflicts of interest con-
Some authors have suggested systematic plating of the non-
cerning this article.
union site with addition of autologous bone, no matter which type
of treatment was initially used [8]. The advantage of this method
is that the radial nerve is inspected, the medullary canal is reper- References
meabilized and rigid fixation is applied to compress the non-union
site, and shorten the humerus as needed. The radial nerve must [1] Graves M, Nork SE. Fractures of the humerus, orthopaedic knowledge update.
Trauma 2010;4:201–24.
be identified some distance from the non-union site and carefully
[2] Coudane H, Bonnevialle P, Bernard J-N, Claudot F. Fractures de la diaphyse
released. humérale chez l’adulte : EMC Appareil locomoteur. Paris: Elsevier Masson SAS;
In cases of non-union following IM nailing or pinning, a larger 2007 [14-039-A-10].
[3] Müller ME, Nazarian S, Koch P, Schatzker J. The comprehensive classification
diameter nail will be used and more importantly, the shaft will be
of fractures of long bones. Berlin: Springer Verlag; 1990.
bored out to stimulate osteogenesis, thereby performing “in situ” [4] Lefevre C. Fractures diaphysaires de l’humérus chez l’adulte. Rev Chir Orthop
autografting. IM nailing can also be performed in cases of hyper- 2004;90(suppl. 5) [1S27–1S67].
trophic non-union. However, Dujardin et al. [31] reported a 38% [5] Bono CM, Grossman MG, Hochwald N, Tornetta P. Radial and axillary nerves:
anatomic considerations for humeral fixation. Clin Orthop 2000;373:259–64.
rate of non-healing after locked IM nail was applied for non-union, [6] Guse TR, Ostrum RF. The surgical anatomy of the radial nerve around the
which could be attributed to lack of stability. humerus. Clin Orthop 1995;320:149–53.
Hybrid external or Ilizarov-type fixation is not without risks [7] Gerwin M, Hotchkiss RN, Weiland AJ. Alternative operative exposure of the
posterior aspect of the humeral diaphysis. J Bone Joint Surg 1996;78A:1690–5.
(nerve damage upon K-wire insertion, stiffness, intolerance to pins, [8] Segonds J-M, Alnot J-Y, Masmejean E. Pseudarthroses et retards de consolida-
elbow septic arthritis) and the fixator must be worn for an average tion aseptiques de la diaphyse humérale. À propos de 30 cas traités par plaque
of 6 months [32]. et autogreffe osseuse. Rev Chir Orthop 2003;89:107–14.
[9] Tytherleig-Strong G, Walls N, Mac Queen MM. The epidemiology of humeral
If radial nerve palsy is associated with the non-union, the shaft fractures. J Bone Joint Surg 1998;80–B:249–53.
bone problem must be treated before the nerve injury so that [10] De Mourgues G, Fischer LP, Gillet JP, Carret JP. Fractures récentes de la diaphyse
the nerve can recovery in an environment with the least amount humérale. Rev Chir Orthop 1975;61:191–207.
 L. Pidhorz / Orthopaedics & Traumatology: Surgery & Research 101 (2015) S41–S49 S49

[11] Sarmiento A, Zagorski JB, Zych GA, Latta LL, Capps CA. Functional bracing [22] Kempf I, Heckel T, Pidhorz LE, Taglang G, Grosse A. L’enclouage verrouillé
for the treatment of fractures of the humeral diaphysis. J Bone Joint Surg selon Seidel des fractures diaphysaires humérales récentes. Rev Chir Orthop
2000;82–A:478–86. 1994;80:5–13.
[12] Nieto H. Les fractures de la diaphyse humérale. Table ronde des jeunes de la [23] Lefevre C. Complications locales et générales des enclouages percutanés :
Société des orthopédistes de l’Ouest. Ann Orthop Ouest 1997;29:129–59. Cahiers d’enseignement de la SOFCOT. Conférences d’enseignement. Expansion
[13] Asencio G, Buscayret F, Trabelsi A, Bertin R, Hammami R, Megy B, et al. scientifique française; 1997. p. 105–19.
Enclouage verrouillé des fractures diaphysaires récentes de l’humérus. Rev Chir [24] Kurup H, Hossain M, Andrew JG. Dynamic compression plating versus locked
Orthop 2001;87:758–64. intramedullary nailing for humeral shaft fractures in adults. Cochrane Database
[14] Paris H, Tropiano P, Clouet D’Orval B, Chaudet H, Poitout D-G. Fractures dia- Syst Rev 2011:15.
physaires de l’humérus : ostéosynthèse systématique par plaque : résultats [25] Butin E, Herent S, Delahaye P. Traitement des fractures de la diaphyse humérale
anatomiques et fonctionnels d’une série de 156 cas et revue de la littérature. par enclouage élastique de Marchetti. À propos de 50 cas. Rev Chir Orthop
Rev Chir Orthop 2000;86:346–59. 2001;87:749–57.
[15] Babin SR. Les fractures de la diaphyse humérale de l’adulte. Cahiers [26] Apard T, Lahogue J-F, Prové S, Hubert L, Talha A, Cronier P, et al. Traitement des
d’enseignement de la SOFCOT. Paris: Expansion scientifique française; 1975. fractures récentes de la diaphyse humérale par enclouage centromédullaire
p. 91–114. verrouillé rétrograde. Rev Chir Orthop 2006;92:19–26.
[16] Sarmiento A, Kinman PB, Galvin EG, Schmitt RH, Phillips JG. Functional [27] Gayet LE, Muller A, Pries P, Merienne JF, Brax P, Soyer J, et al. Fractures de la
bracing of fractures of the shaft of the humerus. J Bone Joint Surg diaphyse humérale : place de l’embrochage fasciculé selon Hackethal. Rev Chir
1977;59–A(5):596–601. orthop 1992;78:13–22.
[17] Dufour O, Beaufils P, Ouaknine M, Vives P, Perreau M. Traitement fonctionnel [28] De La Caffiniere J-Y, Benzimar R, Lacaze F, Chaine A. Embrochage centromé-
des fractures récentes de la diaphyse humérale par la méthode de Sarmiento. dullaire des fractures de la diaphyse humérale. Une ostéosynthèse à risque
Rev Chir Orthop 1989;75:292–300. minimum. À propos de 82 cas. Rev Chir Orthop 1999;85:125–35.
[18] Zagorski JB, Latta LL, Zych GA, Finnieston AR. Diaphyseal fractures of the [29] Alnot J-Y, Osman N, Masmejean E, Wodecki P. Les lésions du nerf radial dans
humerus. Treatment with prefabricated braces. J Bone Joint Surg 1998;70- les fractures de la diaphyse humérale. À propos de 62 cas. Rev Chir Orthop
A(4):607–10. 2000;86:143–50.
[19] Dayez J. Plaque vissée interne dans les fractures récentes de la diaphyse [30] Neuhaus V, Menendez M, Kurylo JC, Dyer GS, Jawa Ring D. Risk factors for frac-
humérale de l’adulte. Rev Chir Orthop 1999;85:238–44. ture mobility six weeks after initiation of brace treatment of mid-diaphyseal
[20] El Ayoubi L, Karmouta A, Roussignol X, Auquit-Auckbur I, Milliez PY, Duparc humeral fractures. J Bone Joint Surg Am 2014;96:403–7.
F. Transposition antérieure du nerf radial dans les fractures du 1/3 moyen [31] Dujardin F-H, Mazirt N, Tobenas A-C, Duparc F, Thomine J-M. Échecs de
de l’humérus : bases anatomiques et applications cliniques. Rev Chir Orthop l’enclouage centromédullaire verrouillé des pseudarthroses de la diaphyse
2003;89:537–43. humérale. Rev Chir Orthop 2000;86:773–80.
[21] Seidel H. Traitement des fractures de l’humérus à l’aide du clou verrouillé. [32] Patel VR, Menon DK, Pool RD, Simonis RB. Non-union of the humerus after
Cahiers d’enseignement de la SOFCOT, 39. Paris: Expansion Scientifique Publi- failure of surgical treatment. Management using the Ilizarov circular fixator. J
cations; 1990. p. 55–9. Bone Joint Surg 2000;82B:977–83.