ORIGINAL ARTICLE
A New Classification System Predictive of Complications
in Surgically Treated Pediatric Humeral Lateral
Condyle Fractures
Jennifer M. Weiss, MD,*w Sara Graves, BA,*w Scott Yang, BS,*w
Elliott Mendelsohn, BS,*w Robert M. Kay, MD,*w and David L. Skaggs, MD*w
Background: The most commonly cited classification system for
lateral condyle fractures (Milch) has not been shown to be
predictive of outcome or recommend treatment.
Purpose: To determine whether a classification system and
treatment based on fracture displacement and articular con-
gruity correlates with the complication rate after pediatric
lateral humeral condyle fractures.
Methods: A retrospective review of all children with lateral
condyle fractures treated operatively at one institution from
1996 to 2003 was performed. All fractures were classified by the
following system: A Type I fracture is displaced less than 2 mm.
In a Type II fracture there is Z2 mm of displacement with intact
articular cartilage, as demonstrated by arthrogram (65 patients).
In a Type III fracture there is Z2 mm of displacement and the
articular surface is not intact (93 patients). The 158 patients with
types 2 and 3 fractures underwent surgery and are the focus
of this study. Complication rates were compared between
groups 2 and 3, and with regard to patient age, length of time
between injury and surgery, and duration of casting.
Results: The overall complication rate was 25% (39 of 158). The
most common complications included radiographic and/or
clinical bump (16 of 158 or 10%), and infection treated with
oral antibiotics (4 of 158 or 2.5%). There were 6% major
complications (10 of 158) defined as those with presumptive
long-term effects or requiring reoperation, including 1 nonunion
(0.6%). There were no acute complications at the time of injury
or surgery. If lateral bump is excluded as a complication, then
the overall complication rate is 14.6% (23 of 158). The overall
complication rates for types 2 and 3 fractures were statistically
significantly different (P<0.03): 11% (7 of 65) for type 2 and
34% (32 of 93) for type 3 fractures. Major complication rates
were 1.5% (1 of 65) for type 2 fractures and 10% (9 of 93) for
type 3 fractures, whereas minor complications occurred in 9%
(6 of 65) of type 2 fractures, and 25% (23 of 93) of type 3
fractures (P = 0.03). There was no correlation between compli-
cation rate and patient age, number of days between fracture
From the *Children’s Orthopaedic Center, Children’s Hospital Los
Angeles; and wKeck-University of Southern California School of
Medicine, Los Angeles, CA.
None of the authors received financial support for this study.
Reprints: Jennifer M. Weiss, MD, Children’s Orthopaedic Center,
4650 Sunset Boulevard, Los Angeles, CA 90027. E-mail: jweiss@
chla.usc.edu.
Copyright r 2009 by Lippincott Williams & Wilkins
602 | www.pedorthopaedics.com
and surgery (all patients were treated within 16 d of their
fracture), or duration of casting. We found that all 65 patients
with Type II fractures had <4 mm of fracture displacement on
pre-operative radiographs, and all fractures Type III fractures
had Z4 mm of displacement. This may aid in predicting
which fractures can be treated with closed pinning prior to an
operative arthrogram.
Conclusions: This is the largest series of operatively treated
lateral condyle fractures reported in the literature. This classi-
fication system and treatment based on fracture displacement
and articular congruity predicts the risk of complications, which
were more than 3 times as likely to occur in type 3 fractures as
type 2 fractures.
Key Words: lateral condyle fracture, children, retrospective
review, pediatric
(J Pediatr Orthop 2009;29:602–605)
L ateral condyle fractures of the distal humerus make
up 12% of elbow fractures in children.1 The most
commonly cited classification system for lateral condyle
fractures (Milch) has not been shown to be predictive of
outcome or recommend treatment.2
When lateral condyle fractures are more than 2 to
3 mm displaced, operative treatment is recommended.3,4
Operative intervention consists of percutaneous fixation
versus open reduction and fixation, depending on the
congruity of the articular surface.5–7 Arthrography has
been shown to correlate well with open operative findings
in lateral condyle fractures.6 When arthrography confirms
congruency of the articular surface, percutaneous pinning
has been shown to be a safe and effective treatment for
lateral condyle fractures.7
Complications of operative fixation for lateral
condyle fractures have been reported to include nonunion,
avascular necrosis, premature epiphysial fusion, lateral
condylar overgrowth, stiffness, and deformity.8–10 Risk
factors for these complications have not been described.
The purpose of this study is to determine whether a
classification system and treatment based on fracture
displacement and articular congruity correlates with a
higher rate of complications in lateral condyle fractures.
J Pediatr Orthop  Volume 29, Number 6, September 2009
J Pediatr Orthop  Volume 29, Number 6, September 2009
MATERIALS AND METHODS
A retrospective review of all children with lateral
condyle fractures treated operatively at one institution
from 1996 to 2003 was performed. All fractures were
classified by the following system: A Type I fracture is
displaced less than 2 mm. In a Type II fracture there is
Z2 mm of displacement with intact articular cartilage, as
demonstrated by arthrogram (65 patients). In a Type III
fracture there is Z2 mm of displacement and the articular
surface is not intact (93 patients). Figure 1 illustrates these
3 types of fractures.
The 158 patients with types 2 and 3 fractures
underwent surgery and are the focus of this study.
Complication rates were compared between groups 2
and 3, and with regard to patient age, length of time
between injury and surgery, and duration of casting.
The surgical technique is as follows:
First, fracture displacement was evaluated. If there
was uncertainty as to the congruity of the articular
surface, an arthrogram was performed. If the articular
surface was congruent, then closed reduction and pinning
were performed. Either 2 or 3 6.2-mm k-wires were placed
in a divergent manner from a lateral entry point. If the
articular surface was not congruent, an open reduction
and internal fixation was performed. The articular surface
was directly visualized and reduced, and either 2 or 3
k-wires were placed in a divergent pattern to stabilize
the fracture. Fluoroscopy was used intraoperatively to
help assess fracture reduction and pin placement. Post-
operatively, the first radiographic assessment was at
1 week after surgery to assure that the fracture reduction
was maintained.
RESULTS
Of the 158 operative lateral condyle fractures
reviewed, 101 patients (64%) were boys, and 57 patients
(36%) were girls. The left side was fractured in 103 of the
elbows (65%) and right side in 55 of the elbows (35%).
FIGURE 1. Type 1 fracture, less than 2 mm displacement;
Type 2 fracture: with Z2 mm displacement and congruity of
the articular surface, and Type 3 fracture: Z2 mm of
displacement and lack of articular congruity. Reproduced with
permission of Children’s Orthopaedic Center, Los Angeles.
r 2009 Lippincott Williams & Wilkins
A Classification System for Lateral Condyle Fractures
There were 39 complications, with a complication
rate of 25%. The overall complication rates for types 2
and 3 fractures were statistically significantly different
(P<0.03): 11% (7 of 65) for type 2 and 34% (32 of 93)
for type 3 fractures.
If a lateral bump is excluded as a complication, then
the overall complication rate is 14.6% (23 of 158.) The
complication rates for types 2 and 3 fractures were not
statistically significant in this scenario (P = 0.15): 12%
(8 of 65) for type 2 and 16% (15 of 93) for type 3 fractures.
There were 10 major complications (10 of 158 or
6%), defined as those with presumptive long-term effects
or requiring reoperation. In the 65 patients with type 2
fractures, there was one major complication (1.5%), a
refracture. The other 9 major complications occurred in
patients with type 3 fractures; 10% of patients with type 3
fractures had a major complication.
There were 29 minor complications (29 of 158 or
18%). Six of these minor complications occurred in
patients with type 2 fractures, 23 occurred in patients with
type 3 fractures.
If lateral bump is excluded as a complication, then
the minor complication rate in type 2 fractures was 1%
(6 of 65) and 0.08% (7 of 93) in type 3 fractures. There is
no statistical significance between complication rates of
types 2 and 3 fractures if lateral bump is excluded as a
complication (P = 0.2). There was one nonunion, which
occurred in a type 3 fracture (0.6%). Complications are
detailed in Table 1.
Of the 65 type 2 fractures that were treated
operatively, 9 (9 of 65 or 14%) were initially type 1
fractures. These displaced despite casting. Of the 65 type 2
fractures, there were 10 complications (15%). Of the 93
type 3 fractures, there were 29 complications (31%).
These complication rates are statistically significantly
different (P<0.03).
There was no correlation between complication rate
and age of the patient. There was no correlation between
the number of days between fracture and surgery and
complication rate. Seventy percent of the children under-
went surgery within 3 days of their fracture. The longest
period of time between fracture and surgery was 16 days.
There was no fracture date available for 11 patients (11 of
158 or 7%).
TABLE 1. Complications
Major Complications
Avascular necrosis 1 0.60%
Malunion 5 3%
Loss of reduction requiring reoperation 1 0.60%
Nonunion requiring reoperation 1 0.60%
Refracture 1 0.60%
Stiffness 1 0.60%
Minor Complications
Infection treated with PO antibiotic 6 3.80%
Keloid 2 1.30%
Radiographic and clinical bump 3 1.90%
Radiographic bump 13 8.20%
Ulcer 1 0.60%
Granuloma 4 2.50%
www.pedorthopaedics.com | 603
Weiss et al
The range of motion data were available for 109
patients. The full range of motion was achieved in 90
(83%) of these children. The limited range of motion was
documented in 19 patients (17%). The average follow-up
for patients with a limited range of motion was 70 days,
ranging from 6 weeks to 6 months. Range of motion did
not correlate with type of fracture. Of the 19 patients with
limited range of motion, 13 sustained type 3 fractures.
There was, however, no statistical correlation between
fracture type and range of motion. The range of motion
did not correlate with casting duration, either. The
average number of days in a cast was 33, and this was
the same for patients with decreased and full range of
motion. Complication rates did not correlate with
duration of casting, either. Patients with complications
averaged 33 days in a cast, as did patients without
complications. Follow-up ranged from 3 weeks to 2 years,
with 3 weeks being the minimum. Radiographic review
did not reveal any complications related to poor pin
placement or configuration.
All 65 patients with Type II fractures had <4 mm
of fracture displacement on pre-operative radiographs
and all fractures Type III fractures had Z4 mm of
displacement. Arthrograms were not performed on patients
with Z4 mm of displacement.
DISCUSSION
The background on other lateral condyle classifica-
tion systems is as follows. The Milch classification defines
a type I fracture as lateral to the trochlear groove. The
fracture may displace, but the elbow does not dislocate.
A Milch type 2 fracture exits medial to the trochlear
groove. The radius and ulna can laterally displace.2
The Jakob classification describes a type 1 fracture as
nondisplaced, a type 2 fracture is complete, and a type 3
fracture has a rotated capitellum.11 The Finnbogason
classification describes a type A fracture as incomplete, a
type B fracture as complete but nondisplaced, and a type
C fracture has a fracture displacement as wide medially as
it is laterally.12 These classification systems were not
designed to dictate treatment or to predict outcome.
The complication rate reported in this study (25%)
falls in the mid-range of rates reported in the literature,
largely because complication rates associated with the
operative treatment of lateral condyle fractures vary
greatly. Mintzer et al7 reported on 12 cases of lateral
condyle fractures with greater than 2 mm of displace-
ment and incongruent artictular surfaces, who underwent
closed reduction with percutaneous pinning. There were
no complications reported among this group of patients.
Mohan et al13 also reported no complications among
20 patients who underwent open reduction with internal
fixation of displaced lateral condyle fractures via postero-
lateral approach.
Higher complication rates are reported by
Ruthorford,14 who saw 10 malreductions and 2 fishtail
deformities among 39 patients (31% complication rate).
Skak et al10 reported a complication rate of 32% with
604 | www.pedorthopaedics.com
J Pediatr Orthop  Volume 29, Number 6, September 2009
1 nonunion, 6 cases of cubitus varus, and 2 cases of
avascular necrosis in 28 patients. Thomas et al15 reported
that late review revealed 61% of children had an
abnormal elbow shape after sustaining a lateral condyle
fracture, which he attributed to overgrowth of the lateral
condyle and excessive formation of bone.
This study examines a classification system and
results of treatment based on the degree of fracture
displacement and presence of articular congruity. Existing
lateral condyle fracture classification systems do not
recommend treatment or predict the outcome.2,11,12,14
Previous classification systems were based on the predic-
tion of displacement (Finnbogason) and sought to predict
displacement of the actual elbow (Milch).2,12 Further-
more, previous work by Mirsky et al16 has described the
limitations of the Milch classification in operative decision
making. The authors cautioned that intraoperative
findings did not correlate with the presumed preoperative
radiographic diagnosis based on the Milch classification
system.
Operatively treated displaced lateral condyle frac-
tures with the articular surface intact (type 2) had half the
complications of displaced fractures without articular
continuity. Whether the higher complication rate in type 3
fractures may be attributable to the articular displace-
ment, increased extent of initial energy, or to the open
reduction is open to speculation.
Complications in this study have been separated
into ‘‘major’’ (those requiring return to the operating
room or resulting in long-term problems) and ‘‘minor.’’
Although the overall complication rate is considerable at
25%, we used a very stringent definition of complications
here, including a radiographic or clinical bump, which
accounts for almost half or 46% (16 of 35) of the
complications. Overall complication rates occurred in
34% of patients with type 3 fractures (32 of 95) and 11%
(7 of 65) of those with type 2 injuries. The likelihood of a
major complication with presumptive long-term effects or
requiring reoperation is only 6% (10% in type 3 and
1.5% in type 2 fractures). This separation is important in
counseling families preoperatively.
One weakness of this study is the short duration of
follow-up. Some complications, including premature
physeal closure, may not be known. Another weakness
of the study is that this classification system is based on
the amount of radiographic displacement. This represents
displacement on the initial radiographs, not the displace-
ment at the time of injury. Thus, true displacement is not
known.
As we found that all 65 patients with Type II
fractures had <4 mm of fracture displacement on pre-
operative radiographs, and all Type III fractures had
Z4 mm of displacement, the question may be asked as to
whether fracture displacement alone can be used for
classification and treatment. However, this 4 mm cut-off
was not a clinical criterion prospectively used for opening
a fracture or doing an arthrogram by the treating
surgeons, but it was rather noted in retrospective review.
In addition, there were only 4 patients with 4 mm of
r 2009 Lippincott Williams & Wilkins
J Pediatr Orthop  Volume 29, Number 6, September 2009
fracture displacement in this series. In this retrospective
series there were no fractures with Z4 mm of displace-
ment in which an arthrogram was performed, so we
cannot claim that all fractures with Z4 mm of displace-
ment are type III fractures with a non-intact articular
surface from our data. Similarly, we cannot draw
conclusions as to the usefulness of arthrograms or the
possibility of closed reduction and pinning in fractures
with Z4 mm of displacement. This may be a worthy topic
for a future prospective study.
In summary, this is the largest series of operatively
treated lateral condyle fractures reported in the literature.
This classification system based on fracture displacement
and articular congruity predicts the risk of complications,
which were more than 3 times as likely to occur in type 3
fractures as type 2 fractures.
REFERENCES
1. Bhandari M, Tornetta P, Swiontkowksi MF. Displaced lateral condyle
fractures of the distal humerus. J Orthop Trauma. 2003;17:306–308.
2. Milch H. Fractures and fracture dislocations of the humeral
condyles. J Trauma. 1964;4:592–607.
3. Flynn JC. Nonunion of slightly displaced fractures of the lateral humeral
condyle in children: an update. J Pediatr Orthop. 1989;9:691–696.
4. Foster DE, Sullivan JA, Gross RH. Lateral humeral condylar
fractures in children. J Pediatr Orthop. 1985;5:16–22.
5. Horn BD, Herman MJ, Crisci K, et al. Fractures of the lateral
humeral condyle: role of the cartilage hinge in fracture stability.
J Pediatr Orthop. 2002;22:8–11.
r 2009 Lippincott Williams & Wilkins
A Classification System for Lateral Condyle Fractures
6. Marzo JM, d’Amato C, Strong M, et al. Usefulness and accuracy of
arthrography in management of lateral humeral condyle fractures in
children. J Pediatr Orthop. 1990;10:317–321.
7. Mintzer CM, Waters PM, Brown DJ, et al. Percutaneous pinning
in the treatment of displaced lateral condyle fractures. J Pediatr
Orthop. 1994;14:462–465.
8. Hasler CC, von Laer L. Prevention of growth disturbances after
fractures of the lateral humeral condyle in children. J Pediatr Orthop B.
2001;10:123–130.
9. Sharma JC, Arora A, Mathur NC, et al. Lateral condylar fractures
of the humerus in children: fixation with partially threaded 4.0-mm
AO cancellous screws. J Trauma. 1995;39:1129–1133.
10. Skak SV, Olsen SD, Smaabrekke A. Deformity after fracture of the
lateral humeral condyle in children. J Pediatr Orthop B. 2001;10:
142–152.
11. Jakob R, Fowles JV, Rang M, et al. Observations concerning
fractures of the lateral humeral condyle in children. J Bone Joint
Surg Br. 1975;57:430–436.
12. Finnbogason T, Karlsson G, Lindberg L, et al. Nondisplaced and
minimally displaced fractures of the lateral humeral condyle in
children: a prospective radiographic investigation of fracture
stability. J Pediatr Orthop. 1995;15:422–425.
13. Mohan N, Hunter JB, Colton CL. The posterolateral approach to
the distal humerus for open reduction and internal fixation of
fractures of the lateral condyle in children. J Bone Joint Surg Br.
2000;82:643–645.
14. Ruthorford A. Fractures of the lateral humeral condyle in children.
J Bone Joint Surg Am. 1985;67:851–856.
15. Thomas DP, Howard AW, Cole WG, et al. Three weeks of
Kirschner wire fixation for displaced lateral condylar fractures of the
humerus in children. J Pediatr Orthop. 2001;21:565–569.
16. Mirsky EC, Karas EH, Weiner LS. Lateral condyle fractures
in children: evaluation of classification and treatment. J Orthop
Trauma. 1997;11:117–120.
www.pedorthopaedics.com | 605