Injury, Int. J.

Care Injured 48 (2017) 2311–2317

Contents lists available at ScienceDirect

Injury
journal homepage: www.elsevier.com/locate/injury

Pilon fractures: A new classification system based on CT-scan

Danilo Leonettia,* , Domenico Tiganib
a
Grande Ospedale Metropolitano, Reggio Calabria, Italy
b
Ospedale Maggiore “CA Pizzardi”, Bologna, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: Actually, pilon fractures are classified according to AO and Ruedi Allgower classification systems based on
Pilon fractures X-rays. These classifications are less reproducible and do not provide necessary information for proper
CT scan classification surgical planning. Aim of the study is to (1) propose a new classification system based on CT scan; (2) to
Treatment-oriented classification
check the prognostic value of this classification and (3) to evaluate its reliability and (4) reproducibility.
Interobserver agreement
We retrospectively reviewed 71 cases of pilon fracture. All fractures were classified according to AO,
Reliability
Ruedi Allgower and new proposed classification system by 5 surgeons. Clinical and radiographic
evaluation were performed at a mean follow-up of 36 months. Cohen’s K value was calculated in order to
evaluate the interobserver and intraobserver agreement.
Sixty-four of 71 fractures healed. Average AOFAS score was 91,7  7,8 in the Type I of new classification
proposed, 87,7  7,8 in the Type II, 82  18,6 in type III, and 67,2  20,9 in type IV. Using the AO
classification system the average K weighted value among the five reviewers was 0,51; using Ruedi
Allgower classification it was 0,50 and using the new classification system it was 0,88 (p < 0.0005).
This study demonstrated that the new classification system is prognostic, reliable and reproducible.
Moreover it provides a new treatment-oriented classification for this challenging fracture which affect
the quality of life of the patients more than chronic diseases like diabetes and coronaropathy or pelvic
fractures.
© 2017 Elsevier Ltd. All rights reserved.

Introduction These classifications are less reproducible and do not provide

adequate information for proper surgical planning [12–14].
The ankle fractures that involve the weight-bearing articular Moreover, these do not provide any correlation between surgical
surface of the distal tibia with proximal extension are known as planning and long term results of the fractures. The aim of this
pilon fractures [1–3]. study is to: (1) propose a new classification system for the pilon
These fractures are among the most complex injuries of the fractures based on the CT scan, useful for the pre-operative surgical
lower limb and their management is technically demanding [4]. planning; (2) check the prognostic value of this classification, and
Surgical treatment of these fractures is challenging, and creates (3) evaluate its reliability and (4) reproducibility.
several difficulties, both in restoration of the articular surface and The study was approved by the local ethics committee and
axial alignment of the tibia and in management of the surrounding follows the Declaration of Helsinki for ethical principles for
soft tissue. medical research involving humans.
In order to obtain good results, a staged management protocol
and a biological friendly surgical technique using a thoughtful CT Materials and methods
scan-based plan are required.
The importance of the pre-operative CT scan evaluation has Classification
been demonstrated in several studies [5–7,9]. Actually, pilon
fractures are classified according to AO [10] and Ruedi Allgower The new classification system is based on 4 labeling factors,
[11] classification systems, which are based only on X-rays. evaluated on the CT-scan: (1) joint involvement of fracture, (2) the
displacement and the number of articular fragments, (3) the plane
along which the major fracture line lies at the joint level, and (4)
areas of comminution.
* Corresponding author.
E-mail address: [email protected] (D. Leonetti).

http://dx.doi.org/10.1016/j.injury.2017.07.026
0020-1383/© 2017 Elsevier Ltd. All rights reserved.
2312 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317

Fig. 1. CT scan classification system of pilon fractures.

According to these findings the classification system is divided Protocol of the study
in four different types (Fig. 1).
Type I are all non-displaced fractures (Type Ia), and all extra- In order to estimate the prognostic value, the reliability and
articular fractures (Type Ib) (Fig. 2). reproducibility of this classification we retrospectively reviewed
Type II are displaced two parts fractures. This group is divided the records of 71 cases of pilon fracture, treated in our Level I
in two subtypes (Fig. 3): Trauma Center from January 2010 to December 2014. Informed
IIS: the fracture line lies on the Sagittal plane (perpendicular to consent was obtained from all individual participants included in
the intermalleolar axis) and separate the pilon in a medial and a the study.
lateral fragment. During this period 96 patients affected by pilon fracture were
IIF: the fracture line lies on the Frontal plane (parallel to the treated in our hospital. Twenty-three patients who presented one
intermalleolar axis) and separate the pilon in an anterior and a of the following criteria were excluded from the study: (1) older
posterior fragment than 75 years; (2) younger than 18 years; (3) lack of preoperative
Type III are displaced three parts fractures. This group is divided CT-scan; (4) neurological disorders. Two patients died before
in two subtype review, and three patients were lost during follow-up.
IIIS: the main fracture line lies on the Sagittal plane At last 68 patients were included in the study, for a total of 71
IIIF: the main fracture line lies on the Frontal plane (Fig. 4). pilon fractures enrolled (3 patients were bilateral) (Table 1).
Type IV are displaced four parts or highly comminuted articular All documents including pre and post-operative clinical reports,
fractures. This type may be divided according to the zone of the X-rays and CT scans were evaluated. Clinical reports were
comminution and contains also the so called “die-punch fragment” examinated in order to check if the inclusion/exclusion criteria
(an articular fragment impacted and raised at metadiaphyseal level were met and to identify operative procedures, associated lesions
as described by Topliss et al. [9]) (Fig. 5). and complications.

Fig. 2. Type I fractures: Ia are non displaced fractures and Ib are extra-articular fractures.
 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317 2313

Fig. 3. Type II fractures: displaced two parts fractures. In IIS fractures (on the left) the fracture line run on the Sagittal plane, in IIF fractures (on the right) the fracture line run
on the Frontal plane.

Fig. 4. Type III fractures: displaced three parts fractures. In IIIS fractures (on the left) the major fracture line run on the Sagittal plane, in IIIF fractures (on the right) the fracture
line run on the Frontal plane.

Fig. 5. Type IV fractures: displaced four parts or comminuted fractures. Comminuted zones are described. In the upper right is shown the Die Punch articular fragment
impacted and raised at metadiaphyseal level.

At last available follow-up, radiographic and clinical evaluation Statistical analysis

using AOFAS score were performed [15]. On X-ray the complete
healing of the fracture and the presence of complications like, The Kendall Tau-B correlation between the AOFAS scores at the
arthrosis, non-union or malunion were assessed. All complications last follow-up and the new classification, the AO classification and
were recorded. the Ruedi Allgower classification, was used to assess the prognostic
All fractures were classified by 5 surgeons accordingly to the AO value of each types of fracture we evaluated. A multivariate
classification, Ruedi Allgower classification and the new system analysis using the multiple regression with backward Wald
proposed. Among the 5 surgeons, one was senior surgeon with method was then performed to detect factors (sex, age, associated
more than 20 years of experience in orthopaedics trauma surgery, lesions, type of surgery, etc.) other than classifications to predict
2 surgeons had more than 5 years of experience in orthopaedics AOFAS. The General Linear Model (GLM), with AOFAS as the
trauma surgery and 2 were senior residents. All the authors dependent variable and the classifications as covariates, was finally
repeated the classification session two times with at least one performed to compare the prediction strength of each classifica-
week between the repeated measurements. tion. The partial eta squared was used to measure the prediction
strength.
2314 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317

Table 1
Clinical details of 71 pilon fractures.

Male: female 51:20 (71–29%)

Average Age at the time of injury 48

(Range; SD) (18–74; 14,6)

Associated injury 29/71; 40%

Surgery
ORIF I stage 9 (12,7%)
ORIF II stage 39 (54,9%)
MIPO 15 (21,1%)
External Fixation 8 (11,3%)

Open fracture
Open/closed fractures (%) 21/50 (29,6%)
Gustilo Anderson 1 7
Gustilo Anderson 2 5
Gustilo Anderson 3a 3
Gustilo Anderson 3b 5 Fig. 6. A graph shows the average Aofas score for each fracture types of new
Gustilo Anderson 3c 1 classification system proposed.

AO classification
A1 3 Table 2
A2 4 Clinical results and complications for each 3 classification systems.
A3 10
Type Cases Average AOFAS Complications
B1 11
B2 10 AO
B3 10 A1 3 79 No
C1 2 A2 4 92 1 Nonunion
C2 7 A3 10 79,7 2 Arthritis
C3 14 1 Malunion
1 Wound Problem
Ruedi Allgower classification B1 11 83 2 Arthritis
I 9 1 Malunion
II 39 B2 10 88,7 1 Malunion
III 23 B3 10 72,2 2 Arthritis
2 Nonunions
New classification system 1 Malunion
I 10 2 Wound Problem
II S 6 C1 2 86 No
II F 10 C2 7 81,7 1 Arthritis
III S 1 1 Malunion
III F 13 1 Infection
IV 31 C3 14 63,1 7 Arthritis
4 Nonunion
1 Malunion
4 Infection
To investigate the reliability of the classification systems we 1 Wound Problem
evaluated the interobserver agreement for each method of
Ruedi Allgower
classification using the weighted Kappa statistics described by
I 9 84,7 1 Arthritis
Fleiss [16,17]. The three Kappa statistics were then compared using II 39 82,8 5 Arthritis
the Wald test. In order to evaluate the reproducibility of the new 2 Infection
proposed classification system we calculated the intra-observer 2 Nonunion
4 Malunion
agreement using the Kappa statistics.
1 Wound Problem
Data analysis was performed using SPSS1 software version 21 III 23 67,8 8 Arthritis
(SPSS Inc., Chicago, IL). The Wald test was performed according to 3 Infection
Shoukri et al. [18]. 5 Nonunion
The literature proposes the following classification for K value: 2 Malunion
3 Wound Problem
less than 0.4 poor agreement, from 0.4 to 0.6 moderate agreement,
from 0.6 to 0.8 good agreement and from 0.8 to 1 excellent New Classification System
agreement [19]. I 10 91,7 1 Nonunion
II 16 87,7 1 Arthritis
1 Infection
Results
1 Wound Problem
III 14 82 2 Arthritis
Sixty-four of 71 fractures healed. The mean follow up was 36 1 Nonunion
months (range 18–71 months). All the fractures are classified 1 Malunion
according AO, Ruedi Allgower and new classification system. IV 31 67,2 11 Arthritis
4 Infections
Average  standard deviation AOFAS score was 91,7  7,8 in the
5 Nonunion
Type I of new classification proposed, 87,7  7,8 in the Type II, 5 Malunion
82  18,6 in type III, and 67,2  20,9 in type IV (Fig. 6). Compli- 3 Wound Problem
cations like arthritis, nonunion, malunion, infection or wound
 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317 2315

Table 3 Interobserver agreement

Prognostic value of 3 different classification systems

Classification system Partial eta squared p-value Using the AO classification system the average K weighted value
New Classification System 0.16 <0.0005 among the five reviewers was 0,51; using Ruedi Allgower
AO 0.057 0.064 classification it was 0,50 and using the new classification system
Ruedi Allgower 0.00001 0.963 it was 0,88 (p < 0.0005). The new classification system showed a
significantly higher inter-observer agreement than the other two
commonly used (p < 0.05).

problems were observed in 29 cases (40,8%). Table 2, summarise Intra-observer agreement

the principal data for type of fracture.
Using the AO classification system the average K weighted value
Prognostic value was 0,59; using the Ruedi Allgower classification was 0,56 and
using the new classification proposed was 0,90 (p < 0.0005).
The new classification showed a higher ordinal correlation with The new classification system has a significantly higher intra-
AOFAS score than AO and Ruedi Allgower classification. observer agreement than the other two classifications (p < 0.05).
Tau B Kendall value was 0.389 for new classification system, No differences were found between experienced and inexperi-
0,293 for AO classification and 0,293 for Ruedi Allgower enced observers.
classification.
Then for each classification, a multivariate analysis using the Discussion
multiple regression with backward Wald method was performed
to detect factors other than the considered classification to predict An ideal fracture classification system should be simple, all
AOFAS. The result of the analysis showed that no other factors inclusive, reliable and reproducible. Furthermore, a treatment
besides the classifications influenced the AOFAS score. Finally, the oriented classification should provide prognostic information
General Linear Model (GLM), with AOFAS as the dependent based on the outcome of different fracture patterns, in order to
variable and the three classifications as covariates, was performed help the surgeon improving preoperative planning and treatment
to compare the prediction strength of each classification, the other [3–12,14].
two being equal. The result was that the new classification was the Currently the tibial pilon fractures are classified according to
significant one and had a higher partial eta squared (Table 3). Ruedi Allgower and AO classification system [10,11].

Fig. 7. Type IIS pilon fracture treated by a medial plate with the screws that intersect the fracture line perpindicularly.
2316 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317

Fig. 8. Type IIIF pilon fracture treated by posterior and medial plate in order to stabilize the posterior and medial fragments.

These systems, are simple, all-inclusive and have a good to moderate reliability and reproducibility [9,12–14,21,22]. This
prognostic value; however they do not provide accurate informa- was thought to be due to difficult identification of displaced
tion on fracture morphology for preoperative planning and articular fragments and the severity of bone injury using plain
demonstrated fair to moderate interobserver reliability [9,12– radiographs. By contrast, the level of intra and inter observer
14,20,21]. agreement with the new classification system was good to
In this study we proposed a new classification system based on excellent.
CT scan in order to be prognostic, reliable, reproducible and useful This difference may be due to the use of a CT scan as a
for preoperative planning and surgical management. preoperative imaging modality, which permits a better identifica-
The new classification system was simple, all inclusive and tion of all articular fragments, their displacements and the areas of
easily to remember. Like Neer classification for proximal humerus comminution. The importance of the CT scan was demonstrated in
fractures, this classification divides the pilon fracture in 4 groups several studies.
according to the number of the fragments at the joint level. Tornetta e Gorup showed that, preoperative CT scan evaluation
Therefore, non-displaced fractures or extra-articular fractures (1 permits to obtain more useful information, like fragments invisible
articular fragment) are included in group I. Two parts fractures are on radiographs, areas of comminution and the direction of the
included in group II, 3 parts fractures in group III and 4 parts and major fracture line in 82% of cases, and caused planning
comminuted fractures are included in group IV. modification and eventually final treatment in 64% of cases [5].
In order to check the prognostic value of the proposed Topliss et al. [9] studied CT scan of 108 cases of pilon fractures,
classification, the correlation between the different types of in order to define the pathological anatomy of these injuries. They
fracture and clinical outcomes were evaluated using the AOFAS described six major articular fragments (anterior, posterior,
scores. The new classification system showed a very good medial, anterolateral, posterolateral and “die-punch”) although
prognostic value with worse clinical results in type III and type not all were universally present. According to the direction of the
IV fractures. major fracture line, 10 types of pilon fracture that could be grouped
Our study demonstrated as good prognostic value as reported in into 2 families (sagittal and coronal) were derived from this study.
the literature for the AO and Ruedi Allgower classifications [21]. According to the aforementioned studies we classified these
However, comparing the prognostic value among all 3 fractures based on the number of articular fragments, their
classification systems, the multivariate analysis showed a higher dislocation, the zones of comminution and the direction of the
significant correlation between clinical results and the new major fracture line. These factors represent a key points for the
classification system if compared to the other classifications. planning and treatment of pilon fractures. In particular, as
The main limitation of currently used classification systems (AO described by Tornetta and Gorup [5], major fracture line direction
and Ruedi Allgower) for pilon fractures is that they have shown fair and zones of comminution determine the incision site and the
 D. Leonetti, D. Tigani / Injury, Int. J. Care Injured 48 (2017) 2311–2317 2317

subsequent internal fixation. Generally if the major fracture line [2] Destot E. Traumatismes du pied et rayons X: malleoles, astragale, calcaneum,
lies on sagittal plane (fracture type IIS and IIIS), is preferred to avant-pied. Paris, 1911, France: Masson; 2017.
[3] Ketz JP, Sanders RW. Pilon fractures. Mann’s surgery of foot and ankle. 9th
apply the fixation device on the medial side of the tibia in order to edition Elsevier Inc.; 2014. p. 1973–2002.
insert screws perpendicular to the fracture line and consequently [4] Blauth M, Bastian L, Krettek C, Knop C, Evans S. Surgical options for the
improve the control on slidings fragments (Fig. 7). If the major treatment of severe tibial pilon fractures: a study of three techniques. J Orthop
Trauma 2001;15(3):153–60.
fracture line runs on frontal plane (fracture type IIF and IIIF), for the [5] Tornetta III P, Gorup J. Axial computed tomography of pilon fractures. Clin
same reason application of the implant fixation on the posterior or Orthop Relat Res 1996;323:273–6.
anterior side of the tibia may be considered (Fig. 8). To achieve the [6] Calori GM, Tagliabue L, Mazza E, de Bellis U, Pierannunzii L, Marelli BM, et al.
Tibial pilon fractures: which method of treatment? Injury 2010;41:1183–90.
maximum stability in the type IV comminuted fractures the [7] Liporace FA, Yoon RS. An adjunct to percutaneous plate insertion to obtain
fixation device should be placed as close as possible to the zone of optimal sagittal plane alignment in the treatment of pilon fractures. J Foot
comminution. Ankle Surg 2012;51:275–7.
[8] Sirkin M, Sanders R, DiPasquale T, Herscovici Jr. D. A staged protocol for soft
tissue management in treatment of complex pilon fractures. J Orthop Trauma
Conclusions 2004;18(Suppl. 8):S32–8.
[9] Topliss CJ, Jackson M, Atkins RM. Anatomy of pilon fractures of the distal tibia. J
In conclusion, the new classification system is prognostic, Bone Joint Surg Br 2005;87(5):692–7.
[10] Müller ME, Nazarian S, Koch P, Schatzker J. The comprehensive classification of
reliable, reproducible and can become an useful instrument to fractures of long bones. Berlin: Springer-Verlag; 1990.
adequately correlate the long term outcomes of these fractures. In [11] Ruedi TP, Allgower M. The operative treatment of intra-articular fractures of
our opinion, the new proposed classification system provides a the lower end of the tibia. Clin Orthop 1979;138:105–10.
[12] Martin JS, Marsh JL, Bonar SK, DeCoster TA, Found EM, Brandser EA.
better three dimensional evaluation of pilon fracture which can Assessment of the AO/ASIF fracture classication for the distal tibia. J Orthop
improve preoperative planning and subsequently surgical treat- Trauma 1997;11:477–83.
ment of this challenging fracture which affect the quality of life [13] Swiontkowski MF, Sands AK, Agel J, Diab M, Schwappach JR, Kreder HJ.
Interobserver variation in the AO/OTA fracture classication system for pilon
more than chronic diseases like diabetes and coronaropathy or fractures: is there a problem? J Orthop Trauma 1997;11:467–70.
pelvic fracture [23]. [14] Dirschl DR, Adams GL. A critical assessment of factors influencing reliability in
The main limitations of this study are the intrinsic weakness of the classification of fractures, using fractures of the tibial plafond as a model. J
Orthop Trauma 1997;11:471–6.
a retrospective study and the great variability of applied surgical
[15] Kitaoka HB, Alexander IJ, Adelaar RS, Nunley JA, Myerson MS, Sanders M.
techniques. Clinical rating systems for the ankle-hindfoot, mid- foot, hallux, and lesser
With the aim to validate this classification system and to use it toes. Foot Ankle Int 1994;15:349–53.
[16] Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas
in a prognostic manner that can drive the surgical plan, a
1960;20(1):37–46.
multicentric/cross cultural prospective study should be designed. [17] Fleiss JL. Measuring nominal scale agreement among many raters. Psychol Bull
1971;76:378–82.
Conflict of interest [18] Shoukri MM, Colak D, Kaya N, Donner A. Comparison of two dependent within
subject coefficients of variation to evaluate the reproducibility of
measurement devices. BMC Med Res Methodol 2008;8:24, doi:http://dx.doi.
The authors declare that they have no conflict of interest. There org/10.1186/1471-2288-8-24.
is no funding source. [19] Landis JR, Koch GG. The measurement of observer agreement for categorical
data. Biometrics 1977;33:159–74.
[20] Cole PA, Mehrle RK, Bhandari M, Zlowodzki M. The pilon map: fracture lines
Acknowledgements and comminution zones in OTA/AO type 43C3 pilon fractures. J Orthop Trauma
2013;27(7):152–6.
[21] Swiontkowski MF, Agel J, McAndrew MP, Burgees AR, MacKenzie EJ. Outcome
The authors are grateful to Dr. Elettra Pignotti for her help and validation of the AO/OTA fracture classification system. J Orthop Trauma
advice on statistics. We are also very grateful to Prof. Alessandro 2000;14:534–41.
Faldini for his incommensurable support and Dr. Mohammadreza [22] Ramappa M, Bajwa A, Singh A, Mackenney P, Hui A, Port A. Interobserver and
intraobserver variations in tibial pilon fracture classification systems. Foot
Cheharassan, Dr. Michele d’Amato, Dr. Paolo Barca for their much 2010;20:61–3.
appreciated work. [23] Pollak AN, Mc Carty ML, Bess RS, Agel J, Swiontkowski MF. Outcomes after
treatment of high energy tibial plafond fractures. J Bone Joint Surg Am
2003;85:1893–900.
References

[1] Michelson J, Moskovitz P, Labropoulos P. The nomenclature for intra-articular

vertical impact fractures of the tibial plafond: pilon versus pylon. Foot Ankle
Int 2004;25:149–50.