International Orthopaedics (SICOT)

DOI 10.1007/s00264-017-3432-3

ORIGINAL PAPER

Comparison of intramedullary nail, plate, and external

fixation in the treatment of distal tibia nonunions
Nabil A. Ebraheim & Brad Evans & Xiaochen Liu & Mina Tanios & Marshall
Gillette & Jiayong Liu1

Received: 12 August 2016 / Accepted: 14 February 2017

# SICOT aisbl 2017

Abstract
Purpose The purpose of this study was to examine time to union of extra-articular distal tibia nonunions based on frac-ture
type and fixation methods: intramedullary nail (IMN), plate osteosynthesis (PO), and external fixation (EF).
Methods This retrospective chart review included all patients who presented at a Level I trauma center with AO/OTA 43A
& distal third 42A-C fracture nonunions between 2008 and 2014. Fixation methods were recorded and patient course was
followed until nonunion had healed clinically.
Results Thirty-three distal tibia nonunions were included, and 29 reached eventual union (88%). Five AO/OTA fracture types were
present. Mean times to union from nonunion diag-
nosis between original fracture types were compared (p = 0.203). Comminuted fracture types had longer times to
union from nonunion diagnosis compared to simple fracture types (78 vs. 46 weeks, p = 0.051) and more revision fixations (1.5 vs.
0.5, p = 0.037). Mean time to union from nonunion
diagnosis was shorter when no revision fixation was done compared to revisions (15 vs. 42 weeks, p = 0.102). Times to union
from nonunion diagnosis without revision fixation
were: IMN (12 weeks), PO (27 weeks), and EF (13 weeks) (p = 0.202). Times to union from definitive revision fixation
were: IMN (17 weeks), PO (21 weeks), and EF (66 weeks) (p = 0.009), with EF taking significantly longer than both oth-er methods.
21 patients (64%) underwent revision fixation. Revision fail rates were: IMN (0/6, 0%), PO (2/8, 25%), and EF (15/21, 71%). Time to
union was longer in revisions that changed fixation method compared to revisions that used the same method (51 vs. 18 weeks, p =
0.030). Deep infections were also associated with longer union times (81 vs. 47 weeks,
p = 0.040).
Conclusions In this nonunion population, comminuted frac-ture types needed more time and revisions to reach union. Time
to union was only clinically shorter when revision fixa-tion was not performed, but IMN and PO were both successful fixation
options with significantly shorter times to union than EF. Mean time to union increased even more when revision of fixation
method was performed vs. exchange revision, as did nonunions with deep infections.

. . . .
Keywords Distal tibia fracture Nonunion Fixation method Fracture type Revision fixation

* Jiayong Liu jiayong.liu@utoledo.edu

1
Department of Orthopaedic Surgery, University of Toledo Medical Center, 3065 Arlington Avenue, Toledo, OH 43614, USA
Introduction

Nonunion of the distal tibia is not uncommon, and they often present with other complications that prolong fracture healing, challenge
orthopaedic surgeons, and have a high morbidity and financial impact [1]. Joveniaux et al. [2] experienced a metaphyseal nonunion
rate of 14%, while Piatkowski et al. [3] experienced a rate of 26.6%. Nonunions are challenging to treat and many surgeons
individualize treatment plans for pa-tients [2, 4, 5]. Known causes of nonunion are medical comor-bidities, tobacco use, and poor
nutrition [6]. Some predictive factors of nonunions include less than 25% cortical continuity, open fractures at injury, wound
complications, comminuted or segmental fractures, need for fibular fixation, inadequate re-duction post-fixation, and infection [7–9].
Revision fixation is often used to correct these gap fractures, and bone grafting is a common and effective technique to stimulate
healing [10].
Non-operative treatment of distal tibia fractures is an option in low energy fractures with acceptable alignment and for
 International Orthopaedics
(SICOT)

patients unfit for surgery; otherwise, surgical fixation is indi-cated. The most common fixation methods are intramedullary nail (IMN),
plate osteosynthesis (PO), and external fixation (EF). Each surgical fixation method has its own advantages and indications [6, 11,
12], so it is difficult to compare methods without also considering the fracture types. Although there are many studies comparing
fixation methods in distal tibia fractures [5, 13–19], treatment superiority has not been established, and these deal with preventing
rather than treating nonunions. Research involved in nonunion man-agement and outcome is either not specific to distal tibia non-
unions [20, 21] or study only one treatment method [22–25]. There is a lack of detailed information involving outcomes of fracture
types and treatment methods of nonunions. This study was designed to evaluate relationships of distal tibia non-unions with fracture
type, initial treatment method, revision failure rate, and definitive treatment methods.

Materials and methods

This is a retrospective review of extra-articular distal tibia nonunions treated at a level 1 trauma center in the United States from 2008
to 2015. This study was approved by the institutional review board. Selection criteria included distal third extra-articular tibia
nonunions (AO/OTA types 43A & distal 42A-C only) [26]. Fractures were considered nonunion after at least six months since injury
with little or no progres-sion toward healing for over three months; 33 patients with nonunion fractures fit these criteria. Information
was gathered from electronic medical records: demographics, medical co-morbidities, radiographs, and course of treatment;
identifying information was removed. Course of treatment was followed until clinical union was achieved: >50% weight-bearing func-
tion with little or no pain, which was at least maintained for three months. AO/OTA guidelines were used to categorize fractures
based on radiographs at the time of injury. Statistical analysis was done with SPSS version 21 [27]. Analysis of variance (ANOVA),
independent t-tests, and uni-variate analysis were done to compare time to union means and measure effect size; tests of
homogeneity of variances and robust tests of equality were done to accommodate heteroge-neous variances and unequal sample
sizes. Significance was set at P < 0.05.

Surgical techniques

Surgeries were performed by several different orthopaedic surgeons and fixation method used was a clinical decision
made by the surgeon. All patients were given IV antibiotics and anaesthesia before surgical fixation. The fractured leg
is then positioned and a tourniquet placed on the proximal thigh before incision.
 For IMN, a medial parapatellar approach is taken with a K-wire guidepin into the tibia. Reaming advancement and
fracture reduction are done simultaneously. The nail is then advanced and fluoroscopy confirms reduction. Distal and
proximal locking screws are then placed.
In PO, approaches are based on fracture and wound loca-tion. Medial, posterolateral, and anterior approaches were used,
with minimal soft tissue stripping toward the bone attempted. Reduction is then accomplished and a locking plate is fitted
across the fracture site. Non-locking screws are used to suck the plate to the bone, and then locking screws are used to fix
position. The area is then thoroughly irrigated and closed.
EF is started with pin placements in the proximal tibia as well as the calcaneous or metatarsals. Large external fixators are
used to hold reduction and then tightened when fluoros-copy confirms satisfactory reduction. Pin sites are then dressed with
Xeroform gauze and wrapped. Ring fixators are included with traditional EF in this study as they have similar indications and
complications.

Results

Thirty-three patients with nonunions met our inclusion criteria. A total of 88% (29) reached union in a mean of 61 weeks from injury
with a mean follow up of 70 weeks (range, 28–240). Two patients were lost to follow up, one patient’s lower limb was amputated due
to nonunion, and one patient died due to an unrelated infection (see Table 1). Only information from the 29 nonunions that eventually
reached union contributed to definitive treatment method and time to union measurements. The average patient age at time of injury
was 49 years (range 22–89), 39% male, 45% obese, 36% current smokers, and 18% with documented dia-betes mellitus. 39% of
the fractures were open and 36% de-veloped deep infections during their treatment course.
In this patient population five distal tibia AO/OTA fracture types were observed: ten simple oblique (42-A2); six simple transverse
(42-A3); two spiral wedge (42-B1); four bending wedge (42-B2); and 11 complex extra-articular (43-A3) (Table 2 shows data based
on fracture type). Mean times to union from nonunion diagnosis of each fracture type were compared (p = 0.203). Due to the sm all
numbers of patients with some fracture types and high variance within each type, no statistically significant differences in time to
union could be attributed to fracture type in this population. Fracture types were then separated based on whether they received
revision fixation (p = 0.699) or not (p = 0.066) (see Chart 1). Fracture types were then categorized as simple or comminuted and
mean times to union from injury of 46 and 78 weeks, respec-tively (p = 0.051, Levene’s test for equality of variances = un-equal
variance in standard deviations). Comminuted fracture
Table 1 Nonunion patient information

Patient Age Male/ BMI Cause of DM Smoker Deep Gustilo type Fracture Initial fixation Revised fixation Bone Bone Weeks to union Weeks to union
number (years) Female injury infection if open type method method grafts grafts from definitive from nonunion
fracture (approach) (approach) (auto) (allo) treatment diagnosis

1 31 F 30.1 - 42-A2 PO (pl) PO (pl) 1 14 3

2 53 M 25.6 Crush x 42-A2 IMN 1 1 39 7
3 49 M 26.8 Fall x x 42-A2 IMN 1 88 24
4 43 F 29.1 Rollerskate 42-A2 IMN 1 32 8
5 47 F 23.0 Fall x 42-A2 IMN IMN 1 11 16
6 39 F 38.8 Fall x x G2 42-A2 IMN 1 62 12
7 44 F 21.9 Motorcycle x 42-A2 EF 1 29 3
8 39 F 31.1 Fall G1 42-A2 EF 1 45 23
9 38 F 40.0 Rollerskate x 42-A2 IMN 32 12
10 43 F 24.8 Long fall x 42-A2 PO (?) IMN 1 14 31
11 87 F 23.6 - 42-A3 PO (m) 1 2 40 14
12 42 F 24.8 Motorcycle G2 42-A3 IMN 1 34 9
13 36 F 25.5 – x 42-A3 IMN IMN 10 10
14 22 M 30.0 – 42-A3 IMN IMN 12 12
15 55 F 36.0 Stress fx x 42-A3 PO (m) PO (am) 1 17 21
16 70 F 24.9 Fall x 42-A3 PO (m) EF x3 1 1 125+ (amputated) 215+ (amputated)
17 89 M 30.8 Fall x x G2 42-B1 IMN EF 1 1 58 48
18 59 M 38.8 Stress fx x G2 42-B1 EF EF 1 1 14+ 14+
19 46 M 22.0 MVA x G2 42-B2 IMN EF 1 51 46
20 52 M 29.3 Fall x 42-B2 IMN IMN 21 22
21 38 M 21.5 GSW x G3A 42-B2 IMN 1 46 12
22 54 F 34.3 Fall x x 42-B2 EF EF x5 30 71
23 65 F 31.6 Fall x G1 43-A3 IMN PO (pl) 2 1 46 50
24 39 M 23.8 Ped v. auto G3B 43-A3 EF PO (pl), EF 2 143 146
25 69 F 33.6 Fall x x 43-A3 EF 1 19+ (died) 54+ (died)
26 43 M 41.3 Fall x G2 43-A3 EF EF 1 33 35
27 45 F 33.3 Osteotomy x x G1 43-A3 PO (m) EF x2 2 3 12+ 178+
28 48 M 25.7 Crush x 43-A3 EF IMN 1 34 11
29 37 F 37.5 Fall 43-A3 PO (m) 1 65 39
30 50 M 27.5 Fall 43-A3 cast PO (a) 14 17
31 43 M 24.3 Motorcycle 43-A3 EF PO (a) 16 7
32 50 F 26.0 MVA x G3B 43-A3 EF PO (al) 1 1 19 9
33 48 F 30.0 Motorcycle x G3B 43-A3 IMN PO (pl), EF x6 2 5 78 205

BMI body mass index, MOI mechanism of injury, DM diabetes mellitus, PO plate osteosynthesis, IMN intramedullary nail, EF external fixation, (N +) indicates a fracture was still
nonunion after N months, (pl) posterolateral, (m) medial, (?) unknown approach, (am) anteromedial, (a) anterior, (al) anterolateral
International Orthopaedics (SICOT)
 International Orthopaedics (SICOT)

Table 2 Outcomes by fracture type

Fracture type 42-A2 42-A3 42-B1 42-B2 43-A3 Overall

Nonunions by initial fracture type 10 6 2 4 11 33

Unhealed nonunions 0 1 1 0 2 4
Nonunions with deep infections 2 (20%) 2 (33%) 1 (50%) 2 (50%) 5 (45%) 12 (36%)
Mean bone grafts (healed nonunions) 1.0 (10) 1.0 (5) 2.0 (2) .5 (2) 1.9 (17) 1.2 (36)
Nonunions that received bone grafts 9/10 3/5 1/1 2/4 7/9 22/29 (76%)
Mean revisions (healed nonunions) 0.3 (3) 0.6 (3) 1.0 (1) 1.8 (7) 1.7 (15) 1.0 (29)
Nonunions that received revisions 3/10 3/5 1/1 3/4 8/9 18/29 (62%)
Revision fail rates (healed & unhealed) 0/3 (0%) 3/6 (50%) 1/2 (50%) 4/7 (57%) 9/17 (53%) 17/35 (49%)
Successful fixations that led to union 1 PO 2 PO 1 EF 2 IMN 5 PO 8 PO
7 IMN 3 IMN 2 EF 1 IMN 13 IMN
2 EF 3 EF 8 EF
Weeks to union from nonunion diagnosis (healed without revision) 13 12 – 12 39 15
Weeks to union from nonunion diagnosis (healed with revision) 17 14 48 46 60 42
Weeks to union from nonunion diagnosis (all 29 healed nonunions) 14 13 48 38 58 32

IMN intramedullary nail, PO plate osteosynthesis, EF external fixation

types also underwent more mean revision fixations at 1.5 vs.

0.5 for simple types (p = 0.037).
A total of 76% (22/29) of patients received a bone graft at some point, with or without a revision fixation and 48% (14/
29) of revisions performed included bone grafting (autograft, allograft, or both). Thirty-eight percent (11/29) of the healed nonunions
reached union with the initial fixation; the remain-ing 62% (18/29) reached union with revision fixation. A total of 29 revisions were
performed on these 18 healed nonunions; 38% of the revisions (11) were considered to have failed. If a revision healed within the
timeframe of this study, but took longer than six months, it was not considered a failure. Revision fail rates include the four nonunions
that still had not reached union if the result was amputation or another revision. Revision fail rates were roughly 50% (50–57%) for
all fracture types, except 42-A2, which was successful in all three revisions; these fracture types also underwent revision fixation less
often. Of note, one 43-A3 fracture underwent seven revisions before finally healing (see Fig. 1). Only three of the 29 nonunions that
eventually reached union underwent more than one revision surgery. Times to union from non-union diagnosis were then grouped
based on whether union
was reached with initial fixation or with revision fixation (15 vs. 42 weeks, p = 0.102). In each fracture type the time to union
increased with revision fixation, not accounting for their varying situations. Union with initial fixation was more com-mon in
simple fracture types and union with revision fixation more common in comminuted types.
IMN was initially placed in 45% (15) of the nonunions, PO on 21% (7), EF on 30% (10), and 3% (1) initially managed non-
operatively due to patient preference (see Table 3). Definitive treatment method is defined as the fixation method present or used
most recently when union is reached, whether
initial hardware or revision. Fixation methods did not heal pro-portionate shares of the fracture types, with most of the plate unions healing 43-A3 type
fractures and most of the IMN unions healing 42-A types, but each treatment method was successful in 43-A3 type nonunions: one IMN, five PO, and
three EF.
When the initial fixation was continued (healed without revision), the mean time to union from nonunion diagnosis was shorter
than with revision fixation from definitive diag-nosis (15 vs. 35 weeks, p = 0.067) (see Chart 2). These two groups were measured
from different time points until union to minimize the effect of any previous failed revision fixation. Initial fixation time to union was
shorter than revision time for all fixation methods except PO. IMN claimed the shortest times to union for both initial and revision (12
vs. 17 weeks, p = 0.256). PO had a longer initial time than revision (27 vs. 21 weeks, p = 0.635); and EF had a very large difference
(13 vs. 66 weeks, p = 0.147). There was no significant difference between mean union times of the different initial treatment methods
(p = 0.212, robust tests of equality = significant vari-ation within each group, partial eta squared = 32% of varia-tions accounted for
by treatment method). There were signif-icant differences among mean union times of revision treat-ment methods (p = 0.009, robust
tests of equality = conflicting results on variance, partial eta squared = 47% of variations accounted for by treatment method).
Revision union times for each method were then directly compared to each other: PO vs. IMN (21 vs. 17 weeks, p = 0.540); IMN vs.
EF (17 vs. 66 weeks, p = 0.020); and PO vs. EF (21 vs. 66 weeks, p = 0.032). EF had significantly longer revision union times
compared to the other fixation methods, while PO and IMN times were not significantly different.
Revision fail rates were lowest in IMN at 0%, then PO at 25%, and then EF at 71%. Although the majority of fractures
International Orthopaedics (SICOT)

Fig. 1 A 43-A3 open fracture in an obese 48-year-old woman in a motorcycle crash. a Initial IMN fixation (L). b IMN removal due to migration, EF placed
(L). c Both locking plate and ring EF revisions done to gradually correct equinus contracture (L). d Union reached with
deformity after almost five years, three more ring fixator revisions and removals, and infection (L). Full strength recovery, limited ankle range
of motion, and minor residual pain

that reached union with IMN were simple fractures, they proved a successful revision fixation method in open frac-tures, infections,
and four of the five fracture types seen in-cluding 43-A3. A total of 66% of nonunions reached union with the same fixation method
used initially: 73% of IMN fractures (11/15); 80% of PO fractures (4/5); and 63% of EF fractures (5/8). This includes exchange
revisions that use the same fixation method (see Fig. 2). In contrast, two of the four nonunions that have not healed have used
multiple fixation methods (see Figs. 3 and 4). When revision of fixation method was deemed necessary, average time to union from
definitive treatment was longer than revisions using the same fixation method (51 vs. 18 weeks, p = 0.030), and time to union from
injury was also longer (86 vs. 53 weeks, p = 0.188).
Deep infections complicated 36% (12/33) of the non-unions: seven infections were managed with EF revision (thre revised from
PO, two from IMN, and two from previous EF); two infections were managed by PO implant removal only; and three were managed
without any surgical intervention other than debridement and drainage. The average time to
union from injury was also much longer in nonunions that developed deep infections compared to those without
(81 vs. 47 weeks, p = 0.040).

Discussion

It is suspected that the lack of surrounding soft tissue, weight bearing function, and a poor blood supply contribute to slower healing
in the distal tibia [4, 6]. The most distal and complex type of these distal tibia nonunions (43-A3) and the other com-minuted types
(42-B1 & B2) experienced longer times to union from injury, longer times to union from nonunion diagnosis, and increased revision
fixations compared to simple nonunions (42-A2 & A3 types). Metsemakers et. al. [28] noticed that in femoral shaft fractures, complex
32C types were more likely to go on to nonunion than simple 32A types. Similarly, this study’s comminuted nonunions had more
complicated courses and more infections in reaching union. This can be interpreted in two ways: more time and revisions are needed
for

Table 3 Outcomes by treatment method

Treatment method IMN PO Ex FIX Overall

Nonunions by initial treatment method 15 (45%) 7 (21%) 10 (30%) 32 + 1 (casted)

Unions by definitive treatment method 13 (39%) 8 (28%) 8 (28%) 29
Revision fail rates (healed & unhealed) 0/6 (0%) 2/8 (25%) 15/21 (71%) 17/35 (49%)
Weeks to union from nonunion diagnosis (healed without revision) 12 weeks (7) 27 weeks (2) 13 weeks (2) 15 weeks (11)
Weeks to union from definitive revision (healed with revision) 17 weeks (6) 21 weeks (6) 66 weeks (6) 35 weeks (17)
Weeks to union from nonunion diagnosis (all 29 healed nonunions) 14 weeks 23 weeks 72 weeks 33 weeks

IMN intramedullary nail, PO plate osteosynthesis, EF external fixation

 International Orthopaedics
(SICOT)

Fig. 2 43-A3 open fracture due to fall in a 65-year-old obese female. a Initial fracture (L). b Initial IMN fixation (L). c IMN removed four months later
due to distal migration. Patient casted for three months then plate placed and bone grafted due to nonunion (L). d After deep
infection with several incisions and drainage procedures the plate was removed. Patient reached union with deformity almost two years after injury,
but suffers chronic pain and chronic osteomyelitis at site

comminuted fractures to eventually reach union, especially be-cause simple fracture types had similar times to union regard-
less of revision fixation; and because orthopaedic surgeons feel uncomfortable with the slower healing rate and attempt to
ex-pedite it with revisions and bone grafts.
Realistic expectations of time to union and need for revi-sion surgery for certain fracture types may help guide or reas-sure
surgeons and patients during fracture healing. It may be comforting that in this study 38% (11/29) healed with no fix-ation revisions,
52% (15/29) underwent just one revision
surgery, and only 10% (3/29) required more than one revision. Of the revisions, 48% included bone grafting, but 60% of bone grafts
were placed without fixation revision. Four pa-tients received bone grafting into an actively infected non-union site, only one of which
healed without any further in-terventions. Sugaya et al. [10] found that a minimally invasive percutaneous bone graft alone was a
safe and effective treat-ment in multiple nonunion sites (femur, tibia, humerus, and ulna) and Braly et al. [29] also found percutaneous
bone graft success specifically in distal tibia metaphyseal plate

Fig. 3 43-A3 fracture due to elective osteotomy in a 45-year-old obese female smoker. a 20° valgus malunion of previous tibial fracture (AP). b Medial plate
placed (AP). c EF Taylor frame placed 12 months later (AP), plate and screw removal due to infection. Failed bone grafts, IM reaming, and fib IMN. d Taylor
frame removed three months later, fracture still
nonunion eight months after removal (AP). e EF placed due to persistent infection and instability (AP). f EF removed four months later,
antibiotic cement placed, another EF placed with bone graft (AP). Patient moved out of the area 3.5 years after osteotomy with EF in place,
resolved infection, but still nonunion
International Orthopaedics (SICOT)

Fig. 4 42-A3 fracture due to a fall in 70-year-old woman with history of vascular insufficiency and previous nonunion 15 years earlier with retained
plate. a Previous plate fixation failure seen and new transverse fracture (AP). b New plate was placed and removed due to abscess
formation, this EF was placed eight months later (AP). c EF revised, then removed, persistent infection (AP). d Taylor frame placed, bone grafting
done, persistent infection (AP). e Taylor frame removed, multiple I&Ds performed. Lower leg amputated five years after injury

nonunions as long as hardware was intact and stable. With the increased time to union with revision fixation in distal tibia
nonunions, a minimally invasive bone graft may be a good starting point for a new nonunion diagnosis.
Fixation method is decided based on many factors, such as fracture type, fracture location, skin injury, and patient comor-bidities.
IMN is typically used in the diaphysis of the tibia, PO in the metaphysis, and EF in severe soft tissue injury and deep infections. Of
our nonunions, the majority were initially treat-ed with IMNs at 45%, EF at 30%, and PO at 21%. This is likely due to a higher number
of distal tibia fractures being treated with IMN rather than lower success of initial IMN fixation; although, we don’t know the fixation
method preva-lence of distal tibia fractures during this period. A meta-analysis done by Zelle et al. [30] looked at over 1100 extra-
articular distal tibia fractures and found 81% were ini-tially treated with IMN and 19% with PO; nonunion rates of each method were
similar at 5.5 and 5.2%, respectively. Fong et al. [7] found the same initial treatment prevalence of 81% IMN and 19% plates when
they examined tibia shaft fractures. Avilucea et al. [5] found that nonunion rates were higher in PO (25%) than IMN (10.6%) in open
distal tibia shaft fractures; but Vallier et al. [14, 15] found higher nonunion rates, infec-tions, and revisions with IMN (7.1%) over PO
(4.2%); a sys-tematic review by Iqbal et al. [13] found infection rates dou-bled in plated fractures; and Sathiyakumar et al. [31] found
higher PO re-operation rates in open vs. closed fractures (46 vs. 29%), confirming the avoidance of PO for open fractures. None of
these studies included EF treatment. If our patient
population experienced a 5% nonunion rate with both IMN and PO as Zelle et al. did, then we could expect the overall
distal tibia fracture population to have similar IMN and PO fixation prevalence as this nonunion population: 68% IMN
and 32% PO (EF excluded).

IMN

Most of the healed nonunions by IMN were in simple aseptic fracture types, which is a common indication [9]. The average time to
union from definitive revision treatments was lowest in IMN at 17 weeks or roughly four months. This number only represents healed
nonunions that underwent a revision IMN, not those that healed with initial hardware, to better compare to other nonunion revision
studies. Other studies have also been successful with IMN revision fixation for distal tibia nonunions with reported times to union
between 3.5 and eight months from revision [21, 32, 33], provided the non-union site allows enough room for the distal screws. IMN
also showed the lowest revision fail rate of 0% compared to 25% of PO and 65% of EF. Successful revisions were done on four of
the five fracture types, including 43-A3 (see Fig. 5).

PO

Our PO time to union from definitive revision was 21 weeks or roughly five months. The only comparable study is Zhao et al. [24],
which treated middle and distal infected nonunions and reported that all infected nonunions healed by 12 months.
 International Orthopaedics
(SICOT)

Fig. 5 42-B2 closed fracture in a 52-year-old male smoker who fell on ice. a Patient was transferred for IMN fixation, fragments of shattered reamer
seen in proximal tibia (AP). b IMN failure due to proximal screw
break and proximal migration six months after injury (AP). c IMN exchange fixation (AP). d Healing fracture, full weight-bearing with minimal
fracture site and knee pain, 13 months after injury

The PO revisions in this study were less successful than IMN with a fail rate of 25% (2/8), both failures being 43-A3 type nonunions
and most of the successes (4/6) also being 43-A3 types (see Fig. 2). The average time to union from definitive revision in 43-A3
types with PO (24 weeks) was much lower than for the IMN (34 weeks) and EF (85 weeks) methods; although, these group sizes
were small. PO was a successful revision method with the shortest times to union in our popu-lation, especially 43-A3 type fractures,
confirming their use in metaphyseal nonunions as well as initial fractures.

EF

Our EF time to union from definitive revision treatment was 66 weeks or roughly 15 months, significantly higher than IMN and PO in
our population (p = 0.009). It is also higher than the 5.8 and 8.8 months reported in other studies [23, 25].
This increased length may be due to the use of EF revision for management of most of the infected nonunions in our popu-lation and
inclusion of traditional external fixators with ring fixators in the EF group. EF experienced a 65% fail rate, which was also higher than
the other methods, but also used as a last resort in this population. All EF revisions were in comminuted fracture types and a single
43-A3 type accounts for five failures and one success (see Fig. 1). Lower fail rates for EF using Ilizarov fixation in the distal tibia
metaphysis have been reported at only 20% by Lonner et al. [22]; al-though, only 1/10 of these fractures was 43-A3 type. Ilizarov
fixation has also proven a successful revision method for infected nonunions: Wu et al. [34] achieved union in all 22 infected PO by
six months, and Megas et al. [35] achieved union with nine infected IMN by an average of six months; but neither of these studies
mentioned fracture types. EF was a less successful revision method in our population compared to

Fig. 6 43-A3 fracture due to a fall in a 43-year-old obese man with diabetes. a Initial Injury (AP). b EF and fib plate placed, K-wires and screws (AP).
EF removed four months after injury; infection develops. Multiple I&Ds and antibiotic cement. c Second EF placed seven months
after injury due to anterior translation of fracture site (AP). d Clinical union reached 15 months after injury; radiographs show healing
(AP); later follow-up confirms radiographic union. Patient has full recovery with mild arthritis
International Orthopaedics (SICOT)

IMN and PO fixation, and was generally avoided as a first-line treatment method.
In this population only 34% (10/29) of the healed non-unions reached union with a fixation method other than the initial fixation
method used. In contrast, two of the four non-unions that have not healed have used more than one fixation method; all of the
revisions being EF. Using a different fixa-tion method increased mean time to union from definitive treatment compared to revision
of the same method (51 vs. 18 weeks, p = 0.030) and time to union from injury (86 vs. 53 weeks, p = 0.188). This increase may be
due to individual fracture/patient factors or due to extra trauma from the method revision itself. This increase may be especially
important for metaphyseal 43-A3 fractures, where the times to union from injury, from nonunion diagnosis, and from definitive treat-
ment are much longer than other fracture types. Vallier et al. [14] speculated their slower IMN healing time was due to disruption of
the medullary blood supply, which could impact healing as a nonunion revision that already has disrupted soft tissue or medullary
blood supply from the current fixation method. Several studies have reported on revisions using the same fixation method to limit
blood and soft tissue disruption. Kruppa et al. [8] reported on 20 metaphyseal IMN nonunions that were treated with nail dynamization
or nail exchange, resulting in successful union of 90% at an average of 13.2 months. Other studies have also reported success with
exchange nailing for aseptic distal tibia nonunions [36, 37]. Revision fixation within the same method (exchange nail, exchange
plating, or exchange external fixation) may provide the stabilization needed without causing the additional trauma of a method revision
(see Figs. 5 and 6).
This study was done as a retrospective review of distal tibia nonunions and was limited by sample size, both in fracture type and fixation methods
used. Not all AO/OTA distal tibia fracture types were represented in this study. Analysis for statistical significance was done with the reservation that
our study had a limited sample size. We found trends in these results, which reveal clinical and some statistical significance but also suggest that a
study with a larger sample size or randomization may have enough power to be more conclusive. More research is needed that compares fracture
types and treatment methods in distal tibia nonunions due to so many non-controlled factors with each nonunion, e.g. surgical approaches, fracture
site and characteristics, and different EF types.

Conclusion

In this nonunion population, comminuted fracture types, es-pecially metaphyseal, needed more time and revisions to reach union.
Deep infections were also associated with prolonged times to union. Time to union was clinically shorter when revision fixation was
not performed, but IMN and PO were
both successful fixation options with significantly shorter times to union than EF (indications excluded). Mean time to union increased
even more when revision of fixation method was performed vs. exchange revision; if revision fixation is deemed necessary to treat a
nonunion, the same fixation meth-od may limit local trauma and allow quicker fracture healing.

Compliance with ethical standards

Conflict of interest All authors declare that they have no conflict of interest.

Funding There is no funding source.

Ethical approval This study was approved by the institutional review board. All procedures performed in studies involving human participants were in accordance
with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable
ethical standards.

Informed consent For this type of study formal consent is not required.

References

1. Tay WH, de Steiger R, Richardson M, Gruen R, Balogh ZJ (2014) Health outcomes of delayed union and nonunion of femoral and tibial
shaft fractures. Injury 45:1653–8
2. Joveniaux P, Ohl X, Harisboure A et al (2010) Distal tibia fractures: management and complications of 101 cases. Int Orthop 34:583–8
3. Piatkowski K, Piekarczyk P, Kwiatkowski K, Przybycien M, Chwedczuk B (2015) Comparison of different locking plate fixa-tion methods in distal
tibia fractures. Int Orthop 39:2245–51
4. Tarkin IS, Siska PA, Zelle BA (2010) Soft tissue and biomechanical challenges encountered with the management of distal tibia non-
unions. Orthop Clin N Am 41:119–26
5. Avilucea FR, Sathiyakumar V, Greenberg SE et al (2015) Open distal tibial shaft fractures: a retrospective comparison of medial plate versus nail fixation.
Eur J Trauma Emerg Surg 42(1):101–106
6. Reed LK, Mormino MA (2008) Distal tibia nonunions. Foot Ankle Clin 13:725–35
7. Fong K, Truong V, Foote CJ et al (2013) Predictors of nonunion and reoperation in patients with fractures of the tibia: an observational study.
BMC Musculoskelet Discord 14:103
8. Kruppa CG, Hoffmann MF, Sietsema DL, Mulder MB, Jones CB (2015) Outcomes after intramedullary nailing of distal tibial
frac-tures. J Orthop Trauma 29:e309–15
9. Tsang ST, Mills LA, Frantzias J, Baren JP, Keating JF, Simpson AH (2016) Exchange nailing for nonunion of diaphyseal fractures of
the tibia: our results and an analysis of the risk factors for failure. Bone Joint J 98-B:534–41
10. Sugaya H, Mishima H, Aoto K et al (2014) Percutaneous autolo-gous concentrated bone marrow grafting in the treatment for
non-union. Eur J Orthop Surg Tr 24:671–8
11. Cabrolier J, Molina M (2015) Is intramedullary nailing superior to plating in patients with extraarticular fracture of the distal tibia?
Med Chem Commun 15:e6306
12. Fadel M, Ahmed MA, Al-Dars AM et al (2015) Ilizarov external fixation versus plate osteosynthesis in the management of extra-articular fractures
of the distal tibia. Int Orthop 39:513–9
 International Orthopaedics
(SICOT)

13. Iqbal HJ, Pidikiti P (2013) Treatment of distal tibia metaphyseal fractures; plating versus intramedullary nailing: a systematic review of recent
evidence. Foot Ankle Surg 19:143–147
14. Vallier HA, Le TT, Bedi A (2008) Radiographic and clinical com-parisons of distal ribia shaft fractures (4 to 11 cm proximal to
the plafond): plating versus intramedullary nailing. J Orthop Trauma 22:207–311
15. Vallier HA, Cureton BA, Patterson BM (2011) Randomized, pro-spective comparison of plate versus intramedullary nail fixation for distal tibia
shaft fractures. J Orthop Trauma 25:736–741
16. Kwok CS, Crossman PT, Loizou CL (2014) Plate versus nail for distal tibial fractures: a systematic review and meta-analysis. J
Orthop Trauma 28:542–8
17. Mao Z, Wang G, Zhang L et al (2015) Intramedullary nailing versus plating for distal tibia fractures without articular involvement: a meta-
analysis. J Orthop Surg Res 10:95
18. Zhang QX, Gao FQ, Sun W, Wang YT, Yang YR, Li Z (2015) Minimally invasive percutaneous plate osteosynthesis versus open
reduction and internal fixation for distal tibial fractures in adults: a meta-analysis. Zhongguo Gu Shang 28:757–62
19. Zhou Y, Wang Y, Liu L, Zhou Z, Cao X (2015) Locking compres-sion plate as an external fixator in the treatment of closed distal
tibial fractures. Int Orthop 39:2227–37
20. Yu Y, Yu KH, Chen Y et al (2014) Comparison of three fixation methods for the treatment of tibial fractures in adolescents.
Zhongguo Gu Shang 27:874–7
21. Shah SB, Mishra AK, Chalise P, Shah RK, Singh RP, Shrivatava MP (2014) Outcome of treatment of nonunion tibial shaft fracture by
intramedullary interlocking nail augmentated with autogenous cancellous bone graft. Nepal Med Coll J 16:58–62
22. Lonner JH, Koval KJ, Golyakhovsky V, Frankel VH (1995) Posttraumatic nonunion of the distal tibial metaphysis-Treatment
using the Ilizarov circular external fixator. Am J Orthop Suppl 16–21
23. Xu K, Fu X, Li YM, Wang CG, Li ZJ (2014) A treatment for large defects of the tibia caused by infected nonunion: ilizarov method with bone
segment extension. Irish J Med Sci 183:423–8
24. Zhao X, Wang PF, Zhang YT, Zhang CC, Xu SG, Zhang X (2014) Advanced bone graft combined with locking compression
plate for the treatment of middle and distal tibia nonunion. Zhongguo Gu Shang 27:1008–11
25. Schoenleber SJ, Hutson JJ Jr (2015) Treatment of hypertrophic distal tibia nonunion and early malunion with callus distraction.
Foot Ankle Int 36:400–7
26. Marsh JL, Slongo TF, Agel J et al (2007) Fracture and dislocation classification compendium - 2007: orthopaedic trauma
association classification, database and outcomes committee. J Orthop Trauma 21(10):S1–S163
27. IBM Corp (2012) IBM SPSS statistics for macintosh, version 21.0. IBM Corp, Armonk
28. Metsemakers WJ, Roels N, Belmans A, Reynders P, Nijs S (2015) Risk factors for nonunion after intramedullary nailing of femoral shaft fractures:
remaining controversies. Injury 46:1601–7
29. Braly HL, O’Connor DP, Brinker MR (2013) Percutaneous autol-ogous bone marrow injection in the treatment of distal meta-
diaphyseal tibial nonunions and delayed unions. J Orthop Trauma 27:527–33
30. Zelle BA, Bhandari M, Espiritu M, Koval KJ, Zlowodzki M (2006) Treatment of distal tibia fractures without articular involvement: a systematic
review of 1125 fractures. J Orthop Trauma 20:76–9
31. Sathiyakumar V, Thakore RV, Ihejirika RC, Obremskey WT, Sethi MK (2014) Distal tibia fractures and medial plating: factors influending
re-operation. Int Orthop 38:1483–8
32. Richmond J, Colleran K, Borens O, Kloen P, Helfet DL (2004) Nonunions of the distal tibia treated by reamed intramedullary
nailing. J Orthop Trauma 18:603–10
33. Paraschou S, Bekir H, Anastasopoulos H et al (2009) Evaluation of interlocking intramedullary nailing in distal tibial fractures and non-unions.
Acta Orthop Traumatol Turc 43:472–7
34. Wu CC (2011) Single-stage surgical treatment of infected nonunion of the distal tibia. J Orthop Trauma 25:156–61
35. Megas P, Saridis A, Kouzelis A, Kallivokas A, Mylonas S, Tyllianakis M (2010) The treatment of infected nonunion of the tibia
following intramedullary nailing by the Ilizarov method. Injury 41:294–9
36. L i t r e n t a J , To r n e t t a P 3 r d , Val l i e r H e t a l ( 2 0 1 5 ) Dynamizations and exchanges: success rates and indications.
J Orthop Trauma 29:569–73
37. Swanson EA, Garrard EC, O’Connor DP, Brinker MR (2015) Results of a systematic approach to exchange nailing for the treat-ment of aseptic
tibial nonunions. J Orthop Trauma 29:28–35