Injury, Int. J.

Care Injured 45 (2014) 667–676

Contents lists available at ScienceDirect

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

Review

Intramedullary nailing versus plating for extra-articular distal tibial

metaphyseal fracture: A systematic review and meta-analysis
Xing-He Xue 1, Shi-Gui Yan, Xun-Zi Cai *, Ming-Min Shi, Tiao Lin
Department of Orthopaedic Surgery, Second Affiliated Hospital, School of Medicine, Zhejiang University, No. 88 Jiefang Road, Hangzhou 310009, PR China

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

Article history: Introduction: With development in the techniques of reduction and fixation, there has been a controversy
Accepted 11 October 2013 in comparison between intramedullary nailing (IMN) and plating for the treatment of distal tibial
metaphyseal fracture (DTF). The study aimed to investigate: (1) which fixation, IMN or plating, was
Keywords: better in the clinical outcomes and in the complications for the treatment of DTF and (2) which
Intramedullary nailing modifying variables affected the comparative results between the two modalities.
Plating Methods: PubMed, EMBASE, OVID, Scopus, ISI Web of Science, the Cochrane Library, Google Scholar and
Distal tibial fracture
specific orthopaedic journals were searched from inception to July 2013, using the search strategy of
Internal fixation
‘(‘Fracture Fixation, Intramedullary’ [MeSH]) AND (‘Tibial Fractures’ [MeSH]) AND (plate OR plating)’. All
prospective and retrospective controlled trials comparing function, pain, bone union and complications
between IMN and plating for DTF were identified. Our analysis had no limitation of the language or the
publication year. The primary outcome measurements were complication rate, union time, operation
time and hospital stays, while the secondary outcome measurements were functional score and pain
score.
Result: Fourteen of 6620 studies with 842 patients were included. IMN was probably preferential to
plating for DTF given its higher functional score (p = 0.01), lower risk of infection (p = 0.02) and
comparable pain score (p = 0.33), total complication rate (p = 0.53) and time to union (p = 0.86).
However, plating had a lower malunion rate than IMN (p < 0.0001). All the results were based on the
Grading of Recommendations Assessment, Development and Evaluation (GRADE) evidence of moderate
quality.
Conclusions: With a satisfying alignment obtained, IMN may be preferential to plating for fixation of DTF
with better function and lower risk of infection. However, IMN showed higher malunion rate for fixation
of DTF. With the biases in our meta-analysis, it will ultimately require a rigorous and adequately
powered randomised controlled trial (RCT) to prove.
Level of evidence: Level III, therapeutic study (systematic review).
ß 2014 Elsevier Ltd. All rights reserved.

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Methodological quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Analysis of the data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Subgroup analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
Statistical analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672
Complication rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672
Functional score and pain score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672
Other data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673

* Corresponding author. Tel.: +86 571 13750838226; fax: +86 571 86806079.
E-mail addresses: qq524689473@163.com (X.-H. Xue), emilcai@hotmail.com (X.-Z. Cai).
1
Tel.: +86 571 15088681606; fax: +86 571 86806079.

0020–1383/$ – see front matter ß 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.injury.2013.10.024
668 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676

Subgroup analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673

GRADE analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676

Introduction 2013), Cochrane Library, Clinical Trial Grade Center and Google
Scholar (1966 to July 2013), Chinese VIP Database (1986 to July
Tibial metaphyseal fractures (TMFs) contain both distal tibial 2013) and Chinese Wan-Fang Database (1992 to July 2013) using
metaphyseal fractures (DTFs, Arbeitsgemeinschaft für Osteo- the search strategy of ‘(‘Fracture Fixation, Intramedullary’ [MeSH])
synthesefragen/Orthopaedic Trauma Association (AO/OTA) type AND (‘Tibial Fractures’ [MeSH]) AND (plate OR plating)’, plus
43 or distal 42) and proximal metaphyseal tibial fractures (PTFs, ‘clinical trial’ AND ‘comparative study’ with no limitation of
AO/OTA type 41 or proximal 42) [1], which account for 3–10% and publication year or language. The reference lists of all the selected
5–11% of total tibial fractures, respectively [2,3]. With the severe articles and the related orthopaedic journals were hand searched
damage of soft tissue and the extreme instability, TMFs have a high for any additional trials. In addition, we searched the Clinical Trial
risk of unsatisfactory function, severe pain, delayed union, Registry, the Current Controlled Trials, the Trials Central and the
malunion and infection [2,4]. The established treatments include Center Watch for grey literature. We defined the criteria of
intramedullary nailing (IMN) and plating. inclusion and exclusion before searching. We only included studies
Plating has been accepted as the first choice for DTF [5], which where: (1) DTF (AO/OTA type 43 or distal 42) was involved, (2) the
ensures accurate reduction and rigid fixation. Unfortunately, age was 18, (3) both IMN and plating were adopted, (4) functional
extensive dissection of the host bone and the soft tissue is score, pain score or complication rate was assessed and (5) the
mandatory. It inevitably raises the risk of infection and nonunion. design was comparative either prospectively or retrospectively.
Furthermore, the complaint about hardware irritation makes it Exclusion criteria included studies where: (1) tibial isthmal
prone to be removed [5]. IMN is the gold standard for tibial fractures or AO type 43-C with serious intra-articular damages
diaphyseal fractures. It has a small influence on the blood supply were involved, (2) neither of the outcomes was available, (3) the
of the host tissue, which would contribute to a low rate of follow-up was <1 year and (4) no control data were provided. All
nonunion and infection [4,5]. Initially, the extreme high malunion the redundant publications were excluded. The abstracts of the
rate and the poor function prevented orthopaedic surgeons from rest of the publications were reviewed for relevance. Excluding
using IMN for DTF [5]. Biomechanical experiments showed that the redundant publications and the unsatisfactory publications,
even the reamed IMN could not fit with the lenient medullary the full texts of the rest of the publications were acquired and read
canal of the tibia metaphysis [6]. Of the two fracture ends, the in detail. We included the publications that satisfied our inclusion
short part lacked the cortical friction with implants and the criteria.
adequate purchase of locking screws so that the tibial alignment We contacted the corresponding authors of the eligible trials if
could be neither obtained nor maintained [6]. Given these necessary to verify the accuracy of the data abstraction as well as
inherent defects, IMNs were limited or even relatively contra- the methodological assessment. We also tried to get any further
dicted for DTF. data or unpublished data which were useful for our data analysis.
With the emerging shortened and multidirectional interlocking
IMN, for example, the distal locking nail (DLN) [2], and the evolving Methodological quality
reduction techniques, for example, the blocking screw (BS) and
other percutaneous reduction techniques (PRTs) [2], the interest in Three reviewers (XHX, SGY and MMS) assessed the methodo-
applying IMN to TMFs has been renewed. The claimed improve- logical quality of the literature according to the 12-item scale [21].
ment in the alignment and the stability has been confirmed both in The 12-item scale contained: randomised adequately, allocation
the laboratory and in the clinical follow-up [2,3,6]. Meanwhile, the concealed, similar baseline, patient blinded, care provider blinded,
novel technique of minimally invasive plating osteosynthesis outcome assessor blinded, avoided selective reporting, similar or
(MIPO) has been developed to further alleviate the local damage of avoided cofactor, patient compliance, acceptable drop-out rate,
plating [3]. similar timing and intention-to-treat (ITT) analysis. The inconsis-
At present, there has been a great controversy on the ideal tent opinions were judged by another author (XZC). The
surgical option for DTF. A large amount of studies compared IMN disagreements were evaluated by the means of a kappa (k) test
with plating [7–20]. Limited by the sample size, they failed to show and resolved by discussion. According to the 12-item standard
a clear superiority of one modality over the other. To address this, (Table 1), five studies [8–12] explicitly described the randomisa-
the present systematic review and meta-analysis is aimed to cover tion and the concealment of the allocation assignment, six studies
all the comparative evidence with the purpose of determining: (1) [8,9,11,14,15,17] described the proper blinding and only one study
which fixation, IMN or plating, was better in the clinical outcomes [9] described ITT analysis. The weighted kappa for the agreement
and in the complications for the treatment of DTF and (2) which on the trial quality between the reviewers was 0.85 (95%
modifying variables affected the comparative results between two confidence interval (CI), 0.77–0.93).
modalities.
Analysis of the data
Materials and methods
Three reviewers (XHX, TL and MMS) extracted the relevant data
Three reviewers (XHX, XZC and MMS) searched PubMed (1966 and checked the accuracy (Table 2). Study design, sample size, age,
to July 2013), EMBASE (1974 to July 2013), Ovid (1966 to July gender, loss to follow-up, AO and Gustilo classification of DTFs,
2013), Scopus (1966 to July 2013), ISI Web of Science (1945 to July reduction technique, implants, fibular fixation, protocol of weight
 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676 669

bearing and outcomes were abstracted. We used the ITT data from

Moderate
Moderate
Moderate

Moderate
Moderate
Moderate
Moderate

Moderate
Qualitye trials whenever it was possible. If the data were not reported in the

High
High

High
High
High

Low
original article, we extrapolated them from the accompanying
graphs. Most of the studies were small scaled with the sample size
ranging from 14 to 160. The total sample size was 443 for IMN and
analysisd

399 for plating. As for IMN, one study [15] adopted the DLN and the
BS, and nine studies attempted the PRTs [8–10,12,14,16,18–20].

Yes
ITT

No
No
No
No
No

No
No
No
No
No
No
No
No
Eight trials [7,10,11,13,15,17,19,20] chose the locking plate and
MIPO was applied in eight others [7–10,15,16,19,20]. The fixation
rates of the associated fibular fracture were >50% in seven studies

Unclear
Similar
timing

[7,11,14,15,17,18,20], <50% in four studies [8,9,13,16] and unclear

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
in three studies [10,12,19]. According to the Gustilo Classification
[22], the fractures in six studies were exclusively closed or Gustilo I
[7,10–14] and the remaining eight also included Gustilo II or
drop-out ratec

Gustilo III open fractures [8,9,15–20].

Only if the method of sequence generated was explicitly described could get a ‘‘Yes’’; sequence generated by ‘‘Dates of Admission’’ or ‘‘Patients Number’’ received a ‘‘No’’.
Acceptable

The primary outcome measurements (Table 3) included the

Unclear
Unclear

Unclear
Unclear
Unclear
Unclear
Unclear

complication rate, the union time, the operation time and the
Yes
Yes

Yes
Yes
Yes

Yes
No

hospital stays. The secondary outcome measurements included

the functional score and the pain score. All the studies mentioned
the malunion and the nonunion of fractures (13/13). The studies
complianceb

showed a higher malunion rate in the IMN group and three

Patient

studies showed a higher nonunion rate in the IMN group. All but
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

one literature mentioned the infection rate (12/13) and only

three studies showed the inferiority of the IMN group. More than
half of the included studies described the secondary surgery rate
Similar or

(7/13) and the union time (8/13). Six studies mentioned the
cofactor
avoided

Unclear

implant removal rate and the operation time (6/13). Few studies
‘‘Yes’’ items greater than 7 means ‘‘High’’; greater than 4 but no more than 7 means ‘‘Moderate’’; no more than 4 means ‘‘Low’’.
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Yes
No

No

mentioned hospital stays (3/13). Functional scores were assessed

ITT = intention-to-treat, only if all randomised patients are analysed in the group they were allocated to could receive a ‘‘Yes’’.

with three different criteria: Olerud and Molander Ankle Score

(OMAS) [11,12,15,17,18] in five studies, American Orthopaedic
reporting
selective
Avoided

Foot and Ankle score (AOFAS) [10] in one and Musculoskeletal

Function Assessment (MFA) [8] in one (7/13 in total). All seven
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

studies showed a higher functional score in the IMN group,

and Mauffery et al. [12] also showed statistical difference
Outcome

(p < 0.05). In addition, four studies assessed pain, of which one

assessor

Unclear

Unclear

Unclear

Unclear

Unclear
Unclear
Unclear
Unclear
Unclear
blinded

assessed knee pain [14], while the other three assessed ankle
Yes

Yes

Yes
Yes

Yes

pain [7,8,10]. As a result, we used the standard mean difference

Intermittent treatment or therapy duration less than 6 months means ‘‘Yes’’, otherwise ‘‘No’’.

(SMD) as the outcome measure of the functional scores and the

pain scores.
provider

Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
blinded
Care

Subgroup analysis
Methodological quality of the included studies based on the 12-items scoring system.

Subgroup analysis was done in our meta-analysis, which

Unclear

Unclear
Unclear

Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
Unclear
blinded
Patient

mainly focuses on the types of internal fixations such as the BS, the
Yes

Yes
Yes
No

locking plate, the DLN and the reamed IMN which would affect the
mechanical stability. The degree of injury of different reduction
techniques, MIPO or non-MIPO, and fracture type made a
baseline

Unclear
Similar

difference. The time of weight bearing also influenced the recovery

of the fracture. In addition, the quality, the study design and the ITT
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Yes

analysis were included in the subgroup analysis.

Drop-out rate > 20% means ‘‘No’’, otherwise ‘‘Yes’’.
Allocation
concealed

Statistical analysis
Yes
Yes

Yes
Yes
Yes
No
No
No

No
No
No
No
No
No

We converted all outcome measurements using Review

Manager 5.1.3 software and Stata 11.0. We used relative risk
(RR) for the dichotomous data and weighted mean difference
Randomised
adequatelya

(WMD) or standardised mean difference (SMD) for the continuous

data. A chi-squared test on N 1 degrees of freedom was used to
calculate the statistical heterogeneity, with significance at 0.05. I2
Yes
Yes

Yes
Yes
Yes
No
No
No

No
No
No
No
No
No

(I2 = ((Q df)/Q)  100%) was used to calculate the percentage of

the variability in effect estimates according to the heterogeneity.
Q is the x2 statistic and df is the degree of freedom. We considered
Mauffery [12]
Janssen [14]

Seyhan [20]
Vallier [16]

I2 values of 25%, 50% and 75% as low, medium and high

Zhang [18]
Vallier [9]
Vallier [8]

Chen [13]
Yang [17]
Guo [10]

Zhu [19]
Feng [7]
Im [11]

heterogeneity, respectively. A fixed-effects model was used if

Li [15]
Study
Table 1

I2 < 50%; otherwise, we used the random-effects model. If

d
b
a

e
c

substantial heterogeneities across studies (I2 > 50%) were

 670
Table 2
Study characteristics and details of interventions of the included studies.

Study Design Sample size Age (years) Gender Follow-up Loss to AO Internal fixation Reduction Fracture Fixed fibula Inclusion
(IMN/plate) (male/ (month) follow up classification (IMN/plate) type (percentage+) criteria
female)

Guo [10] Prospective, 57/54 44.3 (23–70) 50/35 12 13/13 43-A R/MIPO, LCP Percutaneous/ Closed, Gustilo I Unclear Extra-articular
randomised percutaneous (3 cm of distal
fragment)
Im [11] Prospective, 39/39 41.1 (17–65) 46/18 24 5/9 43-A/C1a R/LCP Manual/open Closed, Gustilo I 38 (59%) Extra-articular or
randomised minimally displaced
extension into
the ankle joint
Janssen [14] Retrospective, 12/12 42.1 (25–84) 12/12 63 N.A./N.A. 42-A/B2b R/non-LCP Percutaneous/ Closed, Gustilo I 22 (92%) Extra-articular,
nonrandomised open 18 years, closed
or Gustilo I fracture
Yang [17] Retrospective, 13/14 51.5 (20–86) 12/15 33 N.A./N.A. 43-A Non-R/LCP Manual/open Closed, Gustilo I/II/III 27 (100%) Extra-articular
nonrandomised (3–5 cm of the
distance from the

X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676

fracture line to
ankle articular surface)
Vallier [16] Multicentre, 76/37 39.1 (16–77) 77/34 24 N.A./N.A. 42-Ab R/MIPO, Percutaneous/ Closed, Gustilo I/II/III 42 (36%) Extra-articular
retrospective, non-LCP percutaneous (4–11 cm proximal
nonrandomised from the tibial plafond)
Vallier [9] Prospective, 56/48 38.3 (18–95) 85/19 20 N.A./N.A. 42-Ab R/MIPO, Percutaneous/ Closed, Gustilo I/II/III 28 (27%) Extra-articular
randomised non-LCP percutaneous (4–11 cm proximal
from the tibial plafond)
Vallier [8] Prospective, 56/48 38.3 (18–95) 73/13 22 11/7 42-Ab R/MIPO, Percutaneous/ Closed, Gustilo I/II/III 28 (27%) Extra-articular
randomised non-LCP percutaneous (4–11 cm proximal
from the tibial
plafond)
Mauffery [12] Prospective, 12/12 41.5 (24–60) 8/16 12 N.A./N.A. 43-A Non-R/non-LCP Percutaneous/ Closed, Gustilo I Unclear Extra-articular
randomised percutaneous (within two Müller
squares of the ankle),
>18 years, closed
or Gustilo I
Li [15] Retrospective, 23/23 38.0 (21–59) 36/10 24.7 N.A./N.A. 43-A ETN/MIPO, LCP Manual/manual Closed, Gustilo I/II/III 42 (91%) Closed or Gustilo
nonrandomised I/II fractures,
skeletally mature,
early failure of
conservative
treatment
Seyhan [20] Retrospective, 25/36 39.9 (19–81) 34/27 21.34 N.A./N.A. 43-A R, ETN/MIPO, Percutaneous/ Closed, Gustilo I/II/III 31 (51%) Extra-articular
nonrandomised LCP percutaneous (4–10 cm proximal
from the plafond)
Zhang [18] Retrospective, 27/24 41.2 (18–70) 31/20 21 N.A./N.A. 43-A R/non-LCP Percutaneous/open Closed, Gustilo I/II/III 45 (88%) Extra-articular
nonrandomised
Feng [7] Retrospective, 22/28 44 (19–74) 30/20 12 N.A./N.A. 43-A Non-R/MIPO, Manual/ Closed, Gustilo I 44 (88%) Extra-articular,
nonrandomised LCP percutaneous Gustilo I/II,
Tscherne 0–2
Chen [13] Retrospective, 25/21 31 (21–50) 38/8 24 N.A./N.A. 43-A R/LCP Manual/open Closed, Gustilo I 5 (11%) Extraarticular,
nonrandomised middle and lower
fracture, closed
fracture
Zhu [19] Retrospective, 80/80 38.8 (20–59) 85/62 22.4 6/7 43-A R/MIPO, LCP Percutaneous/open Closed, Gustilo I/II/III Unclear Extraarticular
nonrandomised
IMN, intramedullary nailing; AO, Arbeitsgemeinschaft für Osteosynthesefragen; R, reamed; LCP, locking compression plating; ETN, Expert Tibia Nail; N.A., not applicable; +percentage, number of fixed fibular fracture/total fracture.
a
43-C1 in this study had minimally displaced extension into articular surface which was almost as same as 43-A.
b
All the 42-A cases were 4–11 cm proximal from the tibial plafond.
Table 3
the details of outcomes of included studies.

Study Malunion Nonunion Infection Secondary Implant Time to Operation Hospital Functional Pain
(%) (%) (%) surgery removal union [M(SD)] time [M(SD)] stays [M(SD)] score [M(SD)]a score [M(SD)]b
rate (%) rate (%)

IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate IMN Plate

Guo [10] 0 0 0 0 7 15 52 59 52 59 17.7 17.6 81.23 97.9 N.A. N.A. 86.1 83.9 7.5 8.5
(2.9) (2.2) (11.43) (9.61) (8.16) (6.93) (4.05) (3.91)

X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676

Im [11] 12 0 9 7 3 23 N.A. N.A. N.A. N.A. 18 20 72 89 N.A. N.A. 88.5 88.3 N.A. N.A.
(13) (15) (15) (27.5) (1.76) (1.76)
Janssen [14] 50 17 0 0 8 8 100 75 92 75 21.4 19.3 123 107 9.8 9.5 N.A. N.A. 43.17 6.92
(4.35) (4.97) (30) (33.75) (4) (2.75) (37.91) (14.76)
Yang [17] 23 7 0 0 0 0 N.A. N.A. N.A. N.A. 22.6 27.8 N.A. N.A. 6.4 6.5 86.2 83.9 N.A. N.A.
(4.3) (7.6) (2) (1.6) (3.2) (7.1)
Vallier [16] 29 5 7 3 5 3 30 16 8 8 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.
Vallier [8,9] 25 19 7 4 5 6 18 23 9 10 N.A. N.A. N.A. N.A. N.A. N.A. 73 72 0.35 0.31
(17.25) (16.25) (0.22) (0.24)
Mauffery [12] 8 0 0 8 25 8 8 42 8 33 N.A. N.A. N.A. N.A. N.A. N.A. 82.3 66.7 N.A. N.A.
(13.51) (13.07)
Li [15] 13 4 0 0 4 13 N.A. N.A. N.A. N.A. 21.3 23.1 76 90 5.8 8.9 89 87.6 N.A. N.A.
(3.5) (3.6) (16.6) (20.3) (2.1) (3.1) (7.1) (8.4)
Seyhan [20] 16 3 4 6 0 17 12 56 8 36 15.7 17.24 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.
(3.92) (3.05)
Zhang [18] 19 0 0 0 0 17 N.A. N.A. N.A. N.A. 18.59 (3.75) 20 (3.96) 82.22 (12.51) 89.16 (14.12) N.A. N.A. 82 (7) 79 (7) N.A. N.A.
Feng [7] 23 0 0 0 14 11 N.A. N.A. N.A. N.A. 21.1 15.4 94 100 N.A. N.A. N.A. N.A. 6.7 6.1
(3) (2.9) (12) (14) (2.5) (2.6)
Chen [13] 0 0 0 0 0 0 4 0 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.
Zhu [19] 1 8 0 1 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

IMN, intramedullary nailing; MIPO, minimally invasive plating osteosynthesis; N.A., not applicable; OMAS, olerud and molander functional ankle score; AOFAS, American Orthopaedic Foot and Ankle surgery scores; MFA,
Musculoskeletal Function Assessment.
a
Of the functional score, Guo adopted AOFAS; Im, Yang, Mauffery, Li and Zhang adopted OMAS; Vallier adopted MFA.
b
Of the pain score, Janssen adopted anterior knee pain score; Feng adopted VAS pain score; Guo adopted the pain score of AOFAS; Vallier adopted the pain score of MFA.

671
672 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676
Fig. 1. A QUOROM flowchart illustrated the selection of studies included in our meta-analysis.
detected in the index five main meta-analyses, we performed post
hoc sensitivity analysis by omitting the outlier studies to
determine the sources of heterogeneity. The outliers were
detected as the studies in which the confidence interval of the
estimated effect size did not overlap well with the pooled overall
effect size [23]. For skewed distribution, if the sample size was
<60, standard deviation (SD) = (upper limit lower limit)/4. We
also calculated SD = standard error (SE)  N1/2 if we could got SE
related to Z score. In order to keep the consistent trend in the forest
plot, we used the computational method that adjusted mean
score = total score mean score, and the SD did not change for
several outcomes in the functional score and the pain score. The
funnel plot [24] was used to assess publication bias. If there were
asymmetrical plots, we used the trim and fill analysis to assess the
stability [25]. When allowed, the subgroup analyses were
performed in isolation for DTF. We also used the Grading of
Recommendations Assessment, Development and Evaluation
(GRADE) system to evaluate the quality of evidence by each
outcome [26].
Results
The literature search initially yielded 6620 relevant studies,
from which 2391 redundant publications were excluded. Accord-
ing to our criteria of inclusion and exclusion, 4213 studies were
excluded. In the remaining full texts, three studies had data
duplication and only the newest one was included [7].
One study carried another study further [8,9] and we included
both. Finally, 14 studies including five prospective trials [8–12] and
12 retrospective trials [7,13–20] with 842 participants were
included (Fig. 1). The weighted kappa for the agreement on
eligibility between reviewers was 0.84 (95% CI: 0.71–0.93). No
publication bias was found in the funnel plot (Fig. 2).
Complication rate
IMN decreased the infection rate by 48% (N = 695, RR: 0.52
(0.30, 0.89); p = 0.02), but increased the malunion rate by 147%
(N = 842, RR: 2.47 (1.58, 3.85); p < 0.0001) compared with plating.
There was no significant difference in the total complication rate
(N = 842, RR: 1.14 (0.75, 1.72); p = 0.53), the nonunion rate
(N = 842, RR: 1.16 (0.51, 2.67); p = 0.72), the secondary surgery
rate (N = 457, RR = 0.87 (0.52, 1.43); p = 0.57) and the implant
removal rate (N = 411, RR = 0.82 (0.51, 1.31)) (Table 4). All the
outcomes did not change if the studies with moderate to low
quality were omitted.
We found medium heterogeneity (I2 = 57%) in the total
complication rate. We conducted the subgroup analysis and found
the origin of the heterogeneity. After the separation in the
subgroup analysis of locking plate, the heterogeneity was
significantly reduced (p = 0.04). We also found medium heteroge-
neity in the secondary surgery rate (I2 = 67%) and the implant
removal rate (I2 = 52%). When excluding the data of Seyhan et al.
[23], which included two different IMNs, the heterogeneity
reduced to 37% (p = 0.007) and 16% (p = 0.04), respectively.
Functional score and pain score
IMN increased the functional score (N = 383, SMD: 0.26
( 0.47 to 0.06); p = 0.01) compared with plating. However, no
difference was found in the pain score (N = 245, SMD: 0.23 ( 0.23
to 0.70); p = 0.33). All the outcomes did not change if the studies
with moderate-to-low quality were omitted. We found medium
heterogeneity (I2 = 67%) in the pain score. We noted three studies
[7,8,10] that assessed the ankle pain except for the study by
Janssen et al. [14] that assessed knee pain. The heterogeneity
reduced to 22% after excluding the study by Janssen et al. (p = 0.01).
 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676 673

Fig. 2. Funnel plot for total complication rate between IMN and plating showed no publication bias in visual. IMN, intramedullary nailing; RR, risk ratio.

Table 4
The statistical comparison in outcomes between IMN and plating.

Outcomes Fracture type Event (IMN/plating) Sample size (IMN/plating) Mean [CI] I2 p

1. RR
Total complication rate DTF 118/89 443/399 1.14 [0.75, 1.72] 57% 0.53
Malunion rate DTF 68/22 443/399 2.47 [1.58, 3.85] 29% <0.0001
Nonunion rate DTF 13/9 443/399 1.16 [0.51, 2.67] 0% 0.72
Infection rate DTF 19/35 369/326 0.52 [0.30, 0.89] 15% 0.02
Secondary surgery rate DTF 73/75 250/207 0.87 [0.52, 1.43] 67% 0.57
Implant removal rate DTF 48/58 225/186 0.82 [0.51, 1.31] 53% 0.42
2. WMD
Time to union DTF N.A. 200/208 0.21 [ 2.46, 2.05] 88% 0.86
Operation time DTF N.A. 162/158 10.66 [ 16.64, 4.68] 65% 0.0005
Hospital stays DTF N.A. 48/49 1.08 [ 3.33, 1.17] 79% 0.35
3. SMD
Functional score DTF N.A. 198/185 0.26 [ 0.47, 0.06] 0% 0.01
Pain score DTF N.A. 123/122 0.23 [ 0.23, 0.70] 67% 0.33

DTF, distal tibial fracture; IMN, intramedullary nailing; RR, relative risk; WMD, weighted mean difference; SMD, standardised mean difference; N.A., not applicable; RR > 1 or
WMD > 0 or SMD > 0 means the results favouring plating, vice versa.

All the outcomes did not change if the studies with moderate to (Table 5), there was a reduced trend of PRT in the malunion rate for
low quality were omitted. IMN (p = 0.13). MIPO also showed no statistical significance in the
infection rate compared with IMN (RR: 0.61 (0.30–1.23)). However,
Other data no benefit of MIPO was found in the nonunion rate (p = 0.70).

IMN decreased the operation time (N = 320, WMD = 10.66
( 16.64, 4.68); p = 0.0005) but had no influence on the time to
union (N = 408, WMD: 0.21 ( 2.46 to 2.05); p = 0.86) or the
hospital stays (N = 97, WMD = 1.08 ( 3.33, 1.17); p = 0.35). All
the outcomes did not change if the studies with moderate to low
quality were omitted. The high heterogeneity in the time to union
(I2 = 88%) reduced to 42% (p < 0.00001) after excluding the study
by Feng et al. [7].
Subgroup analysis
IMN conferred more than twice the risk of malunion than
plating. However, if the technique of either the BS or the DLN was
adopted, the inferiority of IMN disappeared (RR of BS: 3.00 (0.34–
26.76), RR of non-BS: 2.45 (1.55–3.85); RR of DLN: 3.00 (0.34–
26.76) and RR of non-DLN: 2.33 (1.46, 3.71)) (Table 5). For DTF
GRADE analysis
Our GRADE analysis showed the comprehensively moderate
quality in all the outcomes (Table 6). The most important reasons
for the reduced level of evidence were inadequate blinding and
lack of concealed allocation. Small sample size also decreased the
evidence grade of the pain score, the union time and the functional
score. Furthermore, the heterogeneity in the pain score and the
union time had a negative effect on the quality.
Discussion
Plating was always accepted as the first choice for DTF until the
late 20th century when IMN gained satisfactory results and wide
popularity with improvement in techniques [2–5]. In this
systematic review and meta-analysis, we asked: (1) which fixation
 674
Table 5
Subgroup analysis of the included studies between IMN and plating based on influential factors.

Factors General complication Malunion rate Nonunion rate Infection rate Secondary operative rate Functional score Pain score

Subgroup RR (95% CI) Subgroup RR (95% CI) Subgroup RR (95% CI) Subgroup RR (95% CI) Subgroup RR (95% CI) Subgroup SMD (95% CI) Subgroup SMD (95% CI)

Designa Pro (4) 1.00 (0.68, 1.47) Pro (4) 1.74 (0.87, 3.47) Pro (4) 1.10 (0.30, 4.06) Pro (4) 0.54 (0.25, 1.13) Pro (3) 0.81 (0.54, 1.22) Pro (4) 0.23 ( 0.48, 0.01) Pro (2) 0.04 ( 0.45, 0.38)
Ret (9) 1.30 (0.69, 2.45) Ret (9) 3.01 (1.68, 5.39) Ret (9) 1.44 (0.61, 3.40) Ret (8) 0.50 (0.23, 1.10) Ret (4) 0.97 (0.36, 2.60) Ret (3) 0.33 ( 0.68, 0.03) Ret (2) 0.66 ( 0.30, 1.62)
p = 0.49 p = 0.23 p = 0.91 p = 0.90 p = 0.75 p = 0.68 p = 0.19

Qualitya Yes (4) 1.00 (0.68, 1.47) Yes (4) 1.74 (0.87, 3.47) Pro (4) 1.10 (0.30, 4.06) Pro (4) 0.54 (0.25, 1.13) Pro (3) 0.81 (0.54, 1.22) Yes (4) 0.23 ( 0.48, 0.01) Yes (2) 0.04 ( 0.45, 0.38)
No (9) 1.37 (0.87, 2.14) No (9) 3.01 (1.68, 5.39) Ret (9) 1.44 (0.61, 3.40) Ret (8) 0.50 (0.23, 1.10) Ret (4) 0.97 (0.36, 2.60) No (3) 0.33 ( 0.68, 0.03) No (2) 0.66 ( 0.30, 1.62)
p = 0.49 p = 0.23 p = 0.91 p = 0.90 p = 0.75 p = 0.68 p = 0.19

ITT Yes (1) 1.29 (0.74, 2.24) Yes (1) 1.33 (0.63, 2.80) Yes (1) 1.71 (0.33, 8.95) Yes (1) 0.86 (0.18, 4.05) Yes (1) 0.78 (0.36, 1.67) Yes (1) 0.66 ( 0.48, 0.36) Yes (1) 0.17 ( 0.25, 0.60)

X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676

No (12) 1.12 (0.69, 1.80) No (12) 3.18 (1.81, 5.56) No (12) 1.02 (0.39, 2.68) No (11) 0.48 (0.27, 0.87) No (6) 0.87 (0.48, 1.57) No (6) 0.32 ( 0.55, 0.09) No (3) 0.31 ( 0.42,1.04)
p = 0.71 p = 0.07 p = 0.59 p = 0.50 p = 0.83 p = 0.28 p = 0.75

ETN Yes (1) 1.00 (0.28, 3.52) Yes (1) 3.00 (0.34,26.76) Yes (1) N.A. Yes (1) 0.33 (0.04, 2.97) N.A N.A. Yes (1) 0.18 ( 0.76, 0.40) Yes (0) N.A.
No (11) 1.27 (0.81, 1.99) No (11) 2.33 (1.46, 3.71) No (11) N.A. No (10) 0.62 (0.35, 1.10) N.A N.A. No (6) 0.27 ( 0.49, 0.06) No (4) N.A.
p = 0.73 p = 0.82 N.A. p = 0.59 N.A. p = 0.76 N.A.

Blocking Yes (1) 1.00 (0.28, 3.52) Yes (1) 3.00 (0.34,26.76) Yes (1) N.A. Yes (1) 0.33 (0.04, 2.97) N.A N.A. Yes (1) 0.18 ( 0.76, 0.40) Yes (0) N.A.
Screw
No (12) 1.15 (0.74, 1.79) No (12) 2.45 (1.55, 3.85) No (12) N.A. No (11) 0.53 (0.31, 0.94) N.A N.A. No (6) 0.27 ( 0.49, 0.06) No (4) N.A.
p = 0.84 p = 0.86 N.A. p = 0.68 N.A. p = 0.76 N.A.

Locking Yes (8) 0.79 (0.46, 1.36) Yes (8) 2.26 (1.10, 4.65) Yes (8) 0.85 (0.25, 2.92) Yes (7) 0.36 (0.17, 0.75) Yes (3) 0.60 (0.17, 2.16) Yes (4) 0.23 ( 0.49, 0.04) Yes (2) 0.04 ( 0.51, 0.42)
plate
No (5) 1.69 (1.05, 2.72) No (5) 2.60 (1.48, 4.57) No (5) 1.49 (0.48, 4.67) No (5) 0.87 (0.38, 2.00) No (4) 1.10 (0.63, 1.92) No (3) 0.31 ( 0.63, 0.00) No (2) 0.62 ( 0.39, 1.63)
p = 0.04 p = 0.77 p = 0.51 p = 0.12 p = 0.24 p = 0.68 p = 0.24

MIPO Yes (7) 1.08 (0.53, 2.21) Yes (7) 2.04 (1.23, 3.39) Yes (7) 1.30 (0.47, 3.60) Yes (6) 0.61 (0.31, 1.18) Yes (4) 0.80 (0.41, 1.53) Yes (3) 0.17 ( 0.44, 0.09) Yes (3) 0.03 ( 0.27, 0.33)
No (6) 1.17 (0.74, 1.85) No (6) 4.43 (1.68, 11.65) No (6) 0.91 (0.22, 3.86) No (6) 0.38 (0.15, 0.98) No (3) 0.84 (0.16, 4.33) No (4) 0.38 ( 0.69, 0.07) No (1) 1.22 (0.33, 2.10)
p = 0.86 p = 0.16 p = 0.70 p = 0.43 p = 0.95 p = 0.31 p = 0.01

Fibular Yes (7) 1.27 (0.73, 2.23) Yes (7) 5.15 (2.31,11.51) Yes (7) 1.06 (0.27, 4.19) Yes (7) 0.30 (0.14, 0.68) N.A. N.A. Yes(1) 0.31 ( 0.73, 0.12) Yes (2) 0.66 ( 0.30, 1.62)
fixationb
No (3) 1.48 (0.55, 3.98) No (3) 2.21 (1.15, 4.24) No (3) 1.99 (0.54, 7.32) No (3) 1.18 (0.34, 4.07) N.A. N.A. No (6) 0.03 ( 0.28, 0.21) No (1) 0.17 ( 0.25, 0.60)
p = 0.80 p = 0.11 p = 0.51 p = 0.07 N.A. p = 0.27 p = 0.36

PRT for Yes (8) 1.11 (0.64, 1.93) Yes (8) 2.05 (1.26, 3.34) Yes (8) 1.12 (0.44, 2.89) Yes (7) 0.58 (0.30, 1.13) Yes (6) 0.84 (0.50, 1.41) Yes (4) 0.30 ( 0.56, 0.05) Yes (3) 0.27 ( 0.38, 0.92)
IMN
No (5) 1.15 (0.59, 2.23) No (5) 5.61 (1.71, 18.46) No (5) 1.32 (0.24, 7.40) No (5) 0.41 (0.16, 1.05) No (1) 2.54 (0.11, 59.23) No (3) 0.19 ( 0.53, 0.15) No (1) 0.23 ( 0.33, 0.79)
p = 0.94 p = 0.13 p = 0.87 p = 0.54 p = 0.50 p = 0.59 p = 0.93

Reamed Yes (9) 1.01 (0.60, 1.69) Yes (9) 2.17 (1.34, 3.52) Yes (9) 1.31 (0.54, 3.14) Yes (8) 0.38 (0.20, 0.75) Yes (6) 0.94 (0.58, 1.53) Yes (4) 0.20 ( 0.44, 0.03) Yes (3) 0.27 ( 0.38, 0.92)
IMN
No (4) 1.50 (0.77, 2.95) No (4) 4.73 (1.43,15.65) No (4) 0.33 (0.01, 7.45) No (4) 1.11 (0.41, 2.98) No (1) 0.20 (0.03, 1.47) No (3) 0.45 ( 0.86, 0.04) No (1) 0.23 ( 0.33, 0.79)
p = 0.36 p = 0.24 p = 0.41 p = 0.08 p = 0.14 p = 0.30 p = 0.93

IMN, intramedullary nailing; ITT, intention-to-treat; ETN, expert tibia nail; MIPO, minimally invasive plating osteosynthesis; PRT, percutaneous reduction technique; Pro, prospective; Ret, retrospective.
a
All the prospective studies were with high quality according to 12-items, while retrospective studies were all with moderate or low quality, yes = high quality according to 12-items scoring system; no = moderate and low quality.
b
Yes = cases with intact fibula or fibular fixation/total cases >50%, no = cases with intact fibula or fibular fixation/total cases <50%.
 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676 675

Table 6
GRADE evidence of comparison between IMN and plating in efficacy and safety for treatment of TMF.

Outcome Summary of findings Quality assessment

Sample size RR/WMD/SMD [95% CI] Limitationsa Inconsistencyb Indirectness Imprecisionc Othersd Quality
(IMN/plating)

Functional score 198/185 0.26 [ 0.47, 0.06] Serious No serious Serious Serious None Moderate
Pain score 123/122 0.23 [ 0.23, 0.70] Serious Serious No serious Serious None Moderate
Time to union 200/208 0.21 [ 2.46, 2.05] Serious Serious No serious No serious None Moderate
Malunion rate 443/399 2.47 [1.58, 3.85] Serious No serious No serious No serious None Moderate
Infection rate 369/326 0.52 [0.30, 0.89] Serious No serious No serious No serious None Moderate
Nonunion rate 443/399 1.16 [0.51, 2.67] Serious No serious No serious No serious None Moderate
Secondary surgery rate 250/207 0.87 [0.52, 1.43] Serious No serious No serious No serious None Moderate
Implant removal rate 225/186 0.82 [0.51, 1.31] Serious No serious No serious No serious None Moderate
Total complication rate 443/399 1.14 [0.75, 1.72] Serious Serious No serious No serious None Moderate
Operation time 162/158 10.66 [ 16.64, 4.68] Serious Serious No serious Serious None Moderate
Hospital stay 48/49 1.08 [ 3.33, 1.17] Serious Serious No serious Serious None Moderate

GRADE, Grading of Recommendations Assessment, Development and Evaluation; RR, risk ratio; WMD, weighted mean difference.
a
Inadequate blinding, lack of allocation concealed may cause limitations.
b
Inconsistent report of outcomes and significant heterogeneity, but we used subgroup analysis to explain them.
c
A study with wide confidence interval around the estimate of the effect, or included sample less than 400, it would cause imprecision.
d
‘‘Other’’ included publication bias and upgraded quality of evidence (large effect, plausible residual confounding and dose-response gradient).

is better with regard to the clinical function and the complications
and (2) which modifying factors affect the comparative effect
between both techniques.
To the best of our knowledge, the present meta-analysis is the
first to comprise all the available comparative controlled evidence
and comprehensively investigate the difference in function, pain
and complications between IMN and plating for DTF. As the
previous systematic reviews only included retrospective observa-
tional studies, their validity was limited by imbalances between
groups, lack of independent assessment, failure for blinding of
outcome measurements and inadequate follow-up [5]. Recently, a
systematic review [27] focussing on complication rate was done by
including two randomised controlled trials (RCTs) [10,11] and four
retrospective comparative studies [14,16,17,28]. All the studies
were also included in our analysis except for one study, which did
not meet our inclusion criteria, involved serious intra-articular
fractures and lacked concrete data of complication rate or
functional score in the IMN group [28]. This systematic review
gave exactly the same results in complication rate, but the lack of
adequate sample size limited the level of evidence. The present
sample size in our analysis is larger with elevated quality of GRADE
evidence compared with the previous reviews. Furthermore, our
analysis has no language restriction so that the publication bias is
reduced as much as possible.
We acknowledge limitations such as: (1) some might argue
against the inclusion of the retrospective studies because of their
inherent risk of bias. However, most of the patient data were
obtained from these studies and, despite methodological limita-
tions, ignoring this source of data might underpower the analysis,
raise the risk of false negative error and influence the accuracy of
our findings. Most of them balanced the demographic parameters
between two groups, which limited the opportunity of selection
bias. Furthermore, the results of the subgroup analysis remained
unchanged after excluding these retrospective studies. (2) The
unavailable raw data of the early studies, such as weight bearing
and fibular fixation, made part of our subgroup analyses
impossible. In addition, the different scoring criteria for function
and pain across the studies might lead to the heterogeneity. These
drawbacks necessitated a uniform and standardised format of
follow-up in future. (3) The heterogeneity was significant in the
outcomes of the total complication rate, the pain score and the
time to union. The sensitivity analysis and the subgroup analysis
were done to find the origins. (4) The small sample size in the
subgroup analysis reduced the precision of the pooled estimates
and the ability to detect the statistical significance of some
variables, that is, the BS, the DLN and the PRTs in the evaluation of
the malunion rate and MIPO in the evaluation of the infection rate.
More RCTs would be warranted to clarify them.
One of the most significant results of our analysis is that IMN
had higher functional score and comparable pain score of DTF
compared with plating. However, we were unable to explain the
clinical implications of the statistical difference because SMD was
merely an absolute value without a unit or a cut-off reference. Of
the included four studies in the pain comparison, three assessed
the ankle [7,8,10], while one assessed the knee [14]. The present
data, indicating the similar rates of hardware removal and the
ankle pain of both modalities, might confirm the possible
relationship between ankle pain and hardware irritation
[7,8,10]. In addition, the present analysis failed to show the
correlation between knee pain and leg function [8,14]. Vallier et al.
[8] found a trend of more knee pain with IMN, whereas 95% of the
patients returned to work without activity limitation. Further, the
more severe knee pain with IMN reported by Janssen et al. [14]
mainly occurred during kneeling or squatting, which did not
influence the excellent knee function.
Our data showed that the malunion rate of IMN was
considerably higher than in the case of plating, which was in
agreement with the biomechanical data [29] and previous
systematic review [27]. Interestingly, when the studies with
combination of IMN and either the BS or the DLN were analysed
separately in the subgroup analysis, the inferiority of IMN
disappeared. In addition, there was a trend of reduced malunion
rate of IMN when the PRTs were adopted. In a multicentre case
series of TMFs treated with the DLN, Attal et al. [30] found that the
malalignment rate was only 5.4% for DTF. Nork et al. [31,32] treated
TMFs using reamed IMN combined with the percutaneous locking
techniques, which showed acceptable alignment in 92% of the
patients. All the clinical and biomechanical evidence supported the
potential advantages of the BS, the DLN and the PRTs.
With regard to DTF, our results demonstrated the excellence of
IMN over plating in the infection rate, which one previous
systematic review failed [5] but another gave the same result
[27]. It turned out that the previous result was potentially biased
by the higher percentage of the open fractures in the IMN group.
Besides, MIPO had an infection rate similar to IMN in the subgroup
analysis, which was consistent with the cadaveric data [33]. An
adequately powered RCT with long-term follow-up is necessary to
verify it.
676 X.-H. Xue et al. / Injury, Int. J. Care Injured 45 (2014) 667–676
The present analysis and the previous systematic reviews [5,27]
both suggested a similar nonunion rate between IMN and plating,
which might be attributed to the minimal exposure, the precise
reduction and the rigid fixation. Despite its theoretical advantage
in local vascularity [33], we did not detect the statistical
significance of MIPO in lowering the nonunion risk. More
comparative data would be needed to address this hypothesis.
Conclusion
IMN and plating appear to share a similar nonunion rate,
secondary surgery rate and implant removal rate for DTF.
Specifically for DTF, IMN may be preferential, given its higher
function score, lower risk of infection and comparable pain score
and time to union. Plating could be an alternative when BS, DLN
and PRT are unavailable. However, with the biases in our meta-
analysis, all these viewpoints will ultimately require a rigorous and
adequately powered RCT to prove.
Conflict of interest
Xun-Zi Cai is currently receiving grants from Zhejiang Provin-
cial Natural Science Foundation of China (Y2110239), National
Natural Science Foundation of China (81101345) and Zhejiang Key
Programme Science and Technology (2011C13033). For the
remaining authors none were declared.
Acknowledgements
We thank the corresponding authors of included studies,
especially Vallier HA, for their sincere assistance in obtaining and
verifying data. The project was funded by Zhejiang Provincial
Natural Science Foundation of China (Y2110239), National Natural
Science Foundation of China (81101345), and Zhejiang Key
Programme Science and Technology (2011C13033).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.injury.2013.10.024.
References
[1] Marsh JL, Slongo TF, Agel J, Broderick JS, Creevey W, DeCoster TA, et al. Fracture
and dislocation classification compendium – 2007: Orthopaedic Trauma As-
sociation Classification, Database and Outcomes Committee. J Orthop Trauma
2007;21:S1–33.
[2] Hiesterman TG, Shafiq BX, Cole PA. Intramedullary nailing of extra-articular
proximal tibia fractures. J Am Acad Orthop Surg 2011;19:690–700.
[3] Bedi A, Le TT, Karunakar MA. Surgical treatment of nonarticular distal tibia
fractures. J Am Acad Orthop Surg 2006;14:406–16.
[4] Bhandari M, Audige L, Ellis T, Hanson B. Operative treatment of extra-articular
proximal tibial fractures. J Orthop Trauma 2003;17:591–5.
[5] Zelle BA, Bhandari M, Espiritu M, Koval KJ, Zlowodzki M. Treatment of distal
tibia fractures without articular involvement: a systematic review of 1125
fractures. J Orthop Trauma 2006;20:76–9.
[6] Bong MR, Kummer FJ, Koval KJ, Egol KA. Intramedullary nailing of the lower
extremity: biomechanics and biology. J Am Acad Orthop Surg 2007;15:97–
106.
[7] Feng YZ, Hong JJ, Peng L, Shui XL, Tang J, Chen LW, et al. Comparison of two
minimally invasive internal fixed methods for the treatment of distal tibio-
fibula. Chin J Surg 2011;2:113–8.
[8] Vallier HA, Cureton BA, Patterson BM. Factors influencing functional outcomes
after distal tibia shaft fractures. J Orthop Trauma 2012;26:178–83.
[9] Vallier HA, Cureton BA, Patterson BM. Randomized, prospective comparison of
plating versus intramedullary nail fixation for distal tibia shaft fractures. J
Orthop Trauma 2011;25:736–41.
[10] Guo JJ, Tang N, Yang HL, Tang TS. A prospective, randomised trial comparing
closed intramedullary nailing with percutaneous plating in the treatment of
distal metaphyseal fractures of the tibia. J Bone Joint Surg Br 2010;92:
984–8.
[11] Im GI, Tae SK. Distal metaphyseal fractures of tibia: a prospective randomized
trial of closed reduction and intramedullary nail versus open reduction and
plating and screws fixation. J Trauma 2005;59:1219–23.
[12] Mauffrey C, McGuinness K, Parsons N, Achten J, Costa ML. A randomised
pilot trial of locking plating fixation versus intramedullary nailing for
extra-articular fractures of the distal tibia. J Bone Joint Surg Br 2012;
94:704–8.
[13] Chen N, He QQ. Clinical analysis for two fixation methods in distal tibiofibular
fracture. Chin Prac Med 2008;3:116–7.
[14] Janssen KW, Biert J, van Kampen A. Treatment of distal tibial fractures: plating
versus nail: a retrospective outcome analysis of matched pairs of patients. Int
Orthop 2007;31:709–14.
[15] Li Y, Liu L, Tang X, Pei F, Wang G, Fang Y, et al. Comparison of low, multidirec-
tional locked nailing and plating in the treatment of distal tibial metadiaphy-
seal fractures. Int Orthop 2012;36:1457–62.
[16] Vallier HA, Le TT, Bedi A. Radiographic and clinical comparisons of distal tibia
shaft fractures (4 to 11 cm proximal to the plafond): plating versus intrame-
dullary nailing. J Orthop Trauma 2008;22:307–11.
[17] Yang SW, Tzeng HM, Chou YJ, Teng HP, Liu HH, Wong CY. Treatment of distal
tibial metaphyseal fractures: plating versus shortened intramedullary nailing.
Injury 2006;37:531–5.
[18] Zhang C, Jiang Y, An ZQ. Interlocking intramedullary nailing versus percuta-
neous plating in osteosynthesis of metaphyseal fractures of distal tibia. Chin J
Orthop Trauma 2007;9:131–4.
[19] Zhu HW. Interlocking intramedullary nail for distal tibial fracture. Chin Prac
Med 2009;4:33.
[20] Seyhan M, Unay K, Sener N. Intramedullary nailing versus percutaneous locked
plating of distal extra-articular tibial fractures: a retrospective study. Eur J
Orthop Surg Traumatol 2012. s00590-012-1016-x.
[21] Furlan AD, Pennick V, Bombardier C, van Tulder M. 2009 updated method
guidelines for systematic reviews in the Cochrane Back Review Group. Spine
(Phila Pa 1976) 2009;34:1929–41.
[22] Fernandez-valencia JA. How to quote Gustilo open fracture classification? J
Shoulder Elbow Surg 2009;18:e32.
[23] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat
Med 2002;21:1539–58.
[24] Higgins JPT, Green S. Cochrane Collaboration Cochrane handbook for system-
atic reviews of interventions. Chichester, England Hoboken NJ: Wiley-Black-
well; 2008.
[25] Hayashino Y, Noguchi Y, Fukui T. Systematic evaluation and comparison of
statistical tests for publication bias. J Epidemiol 2005;15:235–43.
[26] Phi L, Ajaj R, Ramchandani MH, Brant XM, Oluwadara O, Polinovsky O.
Expanding the Grading of Recommendations Assessment, Development,
and Evaluation (Ex-GRADE) for Evidence-Based Clinical Recommendations:
validation study. Open Dent J 2012;6:31–40.
[27] Iqbal HJ, Pidikiti P. Treatment of distal tibia metaphyseal fractures; plating
versus intramedullary nailing: a systematic review of recent evidence. Foot
Ankle Surg 2013 Sep;19(3):143–7.
[28] Joveniaux P, Ohl X, Harisboure A, Berrichi A, Labatut L, Simon P, et al. Distal
tibia fractures: management and complications of 101 cases. Int Orthop
2010;34(4):583–8.
[29] Zlowodzki M, Williamson S, Cole PA, Zardiackas LD, Kregor PJ. Biomechanical
evaluation of the less invasive stabilization system, angled blade plate, and
retrograde intramedullary nail for the internal fixation of distal femur frac-
tures. J Orthop Trauma 2004;18:494–502.
[30] Attal R, Hansen M, Kirjavainen M, Bail H, Hammer TO, Rosenberger R, et al. A
multicentre case series of tibia fractures treated with the Expert Tibia Nail
(ETN). Arch Orthop Trauma Surg 2012;132:975–84.
[31] Nork SE, Schwartz AK, Agel J, Holt SK, Schrick JL, Winquist RA. Intramedullary
nailing of distal metaphyseal tibial fractures. J Bone Joint Surg Am
2005;87:1213–21.
[32] Nork SE, Barei DP, Schildhauer TA, Agel J, Holt SK, Schrick JL, et al. Intrame-
dullary nailing of proximal quarter tibial fractures. J Orthop Trauma 2006;
20:523–8.
[33] Borrelli Jr J, Prickett W, Song E, Becker D, Ricci W. Extraosseous blood supply of
the tibia and the effects of different plating techniques: a human cadaveric
study. J Orthop Trauma 2002;16:691–5.