See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281096595

Fracture non-union epidemiology and treatment

Article  in  Trauma · June 2015

DOI: 10.1177/1460408615591625

CITATIONS READS

40 3,224

3 authors, including:

Rajpal Nandra Liam Grover

Royal Berkshire NHS Foundation Trust University of Birmingham
44 PUBLICATIONS   523 CITATIONS    277 PUBLICATIONS   6,194 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Biomaterials for cartilage and bone repair in osteoarthritis View project

A systematic review of objective burn scar measurements View project

All content following this page was uploaded by Rajpal Nandra on 27 March 2018.

The user has requested enhancement of the downloaded file.

 TRAUMA
Review Article
Trauma
0(0) 1–9

Fracture non-union epidemiology and ! The Author(s) 2015

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
treatment DOI: 10.1177/1460408615591625
tra.sagepub.com

Rajpal Nandra, Liam Grover and Keith Porter

Abstract
Fracture non-union remains a clinical problem despite advances in the understanding of basic science and technology.
Each fracture has a unique personality as does the patient suffering the injury. Thus, each case must be treated on an
individual basis. This article defines the problem of fracture non-union and reports recent epidemiological studies. We
discuss relevant risk factors and methods for assessing patients who have a tendency toward fracture non-union. There
are many treatment options for patients with non-union, where a number of these modalities are still under review. We
discuss current evidence with the use of bone morphogenic protein, platelet-rich plasma and low-intensity pulsed
ultrasound to augment the treatment of fracture non-union.

Keywords
Non-union, incidence, risk factors, platelet-rich plasma, low-intensity pulsed ultrasound

the patient factors and fracture characteristic which

Introduction
correlate with a higher incidence of non-union, describe
Orthopaedic practice in the management of acute frac- the commonly used Weber and Cech (1976) classifica-
tures has been redefined through a greater understand- tion used to facilitate management and discuss the
ing of fracture biology and soft tissue injury, allied to merits and pitfalls of common treatments.
innovative implants, improved operative techniques
and enhanced rehabilitation. Despite these advances
in care, the incidence of fracture non-union remains a
Definition
concern. Non-union is an outcome that is difficult to A non-union exists when repair is not complete within
predict at the time of injury and during the healing the period expected for a specific fracture and when
process and although an established complication of cellular activity at the fracture site ceases and there is
any fracture, in clinical practice both clinicians and no visible progressive signs of healing for 3 months.1,2
patients deem it a poor outcome. There may be persist- The healing process prematurely terminates with no
ent pain and failure to return to pre-injury levels of further radiological evidence of consolidation. With
function. varying repair times for individual fractures and heal-
Fracture non-union is broadly divided into hyper- ing potential of patients, there is no uniform definition
troophic non-union (callus formation) or atrophic for non-union. According to American Food and Drug
non-union (no callus and bone resorption) with a Administration, a non-union is established when a min-
multi-faceted aetiology. Non-union occurs when there imum of 9 months has elapsed since injury and the
is failure of the biology (high-energy injuries with fracture shows no visible progressive signs of healing
devascularisation), failure of the host (nicotine con- for three months.3 It is impractical to apply this to
sumption, vascular disease, other comorbidities), fail- every patient, and clinician’s modify according to
ure of mechanics (improper stabilisation) or treatment each individual scenario. Delayed union, non-union
failure (iatrogenic devascularisation).1 If there has been
any previous surgery to the limb, there should always Queen Elizabeth Hospital, Birmingham, UK
be the suspicion of infection as the cause of non-union.1 Corresponding author:
This article discusses the true extent of fracture non- Rajpal Nandra, Queen Elizabeth Hospital, Birmingham B152TH, UK.
union in adult and paediatric populations. We report Email: rajpalnandra@nhs.net

Downloaded from tra.sagepub.com by guest on June 25, 2015

2 Trauma 0(0)

and pseudoarthrosis represent part of the spectrum of adults, forearm non-union is most common, and with
bone repair. It is difficult to predict which patient or increasing age there is a shift to proximal long bones
fracture type will progress to non-union and an even (humerus and femur). This reflects the population stu-
greater challenge is preventing it. died and overall incidence of fractures. Interestingly,
the non-union rate is higher in females during the
fifth decade, attributable to fragility fractures.
Incidence
The reported incidence and prevalence of non-union
vary significantly based on anatomic region and the
Paediatric fractures
criteria used to define non-union. It has been estimated Non-union in children is considered a rare complication.
that 100,000 fractures go on to non-union each year The immature skeleton has a robust periosteal layer, suf-
in the USA;4 however, prior to epidemiological studies ficient vascularity and enhanced healing potential. Over
reported in 2013, there is a paucity of population-based a 5-year period (2005–2010) in Scotland, there were 180
data, and non-union rates have traditionally been children under the age of 15 years treated for non-union
fracture- or treatment-specific.5 Table 1 reports the (60.6% in males, 39.4% in females), and in the same
non union rates for common long bone fractures and period there were 424 treated fracture non-union
varying treatment methods. With little research in this patients between the ages of 15 and 19 years (71.9%
field, it raises the question whether we appreciate the and 28.1% males and females, respectively, in the
true magnitude of non-union in clinical practice and the under-20 age group).6 The risk of non-union per fracture
substantial direct and indirect socioeconomic costs. is low throughout childhood with a risk of approxi-
mately 1 in 500 or less per fracture in boys aged under
14 years and in girls of all ages; however, non-union risk
Adult fractures increases to approximately 1 in 200 fractures for the
In the UK, there are approximately 850,000 new adult older teenage (15–19 years) boys.6
fractures seen each year.5 A retrospective analysis of
Scottish admission coding databases from 2005 to
2010 found an average of reported 979 new cases of
Socioeconomic cost
non-union in Scotland and found an overall incidence Fracture non-union has direct cost implication to the
of non-union to be 18.4/100,000 per annum, with a National Health Service and an indirect cost to each
peak at 30 to 40 years of age, and more common in patient through loss of earnings and additional social
males.5 Scrutinising their data, it appears that the inci- care. The indirect costs can be large considering the pro-
dence in females has remained static over five years, portion of working aged individuals who are affected. For
and in contrast there has been a 10% reduction in this reason, it has been suggested that waiting for 9
males. Male incidence is bimodal, which mirrors the months for a diagnosis may be unrealistic before interven-
incidence of fractures overall peaking at 25 to 30 tion.8 The direct cost of treating non-union has been
years and then again at 80 years of age. In young reported to range between 7000 and 79,000 per

Table 1. Non-union rates in specific long bone shaft fractures. The specific rates are listed according to the different treatment
modalities available.5,7

Non-union
Fracture Why Treatment rate

Humeral Shaft 1–3% of all fractures Conservative 4.2%

Debate over conservative vs. operative man- Plate fixation 4.2%
agement with plate fixation Antegrade IM nail 11.6%
Retrograde IM nail 4.5%
Femoral Shaft Requires stable fixation, early mobilisation and Plate osteosynthesis 2 to 7%
union. Transition from plate and biological Antegrade IM nail 1.1%
plate to IM fixation Retrograde IM nail 5.8%
Tibial Shaft 26/100,000 year Conservative þ cast 0 to 5%
High energy, associated soft tissue injury with Extra medullary plate 1%
high non-union rate and morbidity Reamed nail 1.6–8%
Unreamed 11%

Downloaded from tra.sagepub.com by guest on June 25, 2015

Nandra et al. 3

person.5 Average direct costs of treatment for an estab- appearance (Elephant foot, Horse hoof and
lished long bone non-union have been reported as follows: oligotrophic).
Canada, $11,800; the USA $11,333; the UK 29,204.4 Avascular non-unions lack vascularity and biological
In 2007, Kanakaris and Giannoudis investigated the healing potential and show no evidence of healing and
economic cost to hospital organisations for individual are associated with factors acting directly on the early
patients with non-union. During the financial year of phases of fracture healing, while hypertrophic non-
2006–2007, the authors collected costs for beds, unions relate mostly with factors acting on the ‘re-orga-
implants, staffing, and investigations and calculated the nisation’ phase of bone healing.15 Avascular non-union
costs by devising a best-case scenario treatment pathway can be further sub-grouped by the fracture pattern: tor-
for each fracture. This included a single corrective surgi- sion wedge, comminuted, defect and atrophic.11
cal procedure ultimately leading to successful union. For The classification systems to date characterise estab-
a humeral fracture non-union, treated with compression lished non-union cases and are useful in treatment plan-
plate fixation and grafting, with a length of in-hospital ning. It is important to consider the possibility of infection
stay of four days, four outpatient clinic visits and mean in all cases, particularly in high-energy open fractures.
time to union of four months, the best case scenario cost
was approximately 15,566. In a best case scenario of
femoral or tibial fracture non-union with exchange nail-
Risk factors
ing, the cost was estimated at 17,200 and 16,330, Many non-unions are associated with soft tissue damage,
respectively.9 Indirect costs are difficult to quantify, as fracture site avascularity, persistent instability or infec-
each patient’s circumstances, job and earnings differ. tion. The etiology is more complicated than this, with a
Hak et al. in 2014 claims indirect costs of 67–79% of host of patient and treatment factors being influential to
the total costs of a tibia fracture in the Canadian each case. Table 2 summarises the common factors asso-
health care system and 82.8–93.3% of all costs for tibia ciated with non-union in the literature to date. These are
fracture patients within European health care systems.4 grouped into patient and fracture characteristics.
A study of litigation claims over 15 years revealed We have discussed earlier how the young adult
further financial burden to health authorities. population is prone to upper limb non-unions. In this
Litigation procedures are increasing, particularly when young working population, it would be useful to pre-
there is associated residual deformity. 143 patients with dict the risk of non-union on an individual basis to
lower limb non-union payouts amounted to 5.3 million counsel and modify treatment accordingly. Perhaps fac-
(mean 75,866 per person). 60% of claims received com- tors such as smoking and dietary insufficiency can be
pensation with mean legal fees of 24,680 per patient.10 modified to minimise risk.
In an economic climate where healthcare systems are The local environment, at the fracture site, is crucial
under strain, this is a growing concern. Patient education to progression of healing. Callus formation is a function
and expectations must be addressed in conjunction with of many variables which we can alter as clinicians. It is
advancement in techniques and surgeon competence to known that callus is a function of both the magnitude
reduce this financial burden. and frequency of inter-fragmentary motion.14 Fracture
distraction more than 3 mm during stabilisation of tibial
fractures increases the odds of non-union by four-fold.14
Classification Inadequate reduction leading to large fracture gaps par-
Weber introduced a classification system based on bio- allels non-union rates in proximal femoral fractures.16,17
logical activity at the fracture site which also facilitates Treatment must focus on avoiding poor anatomical frac-
treatment decisions.11 It is widely used and has stood ture reduction, excess periosteal stripping and prolonged
the test of time. Other classifications exist, that are spe- immobilisation.
cific to a particular fracture site or injury type.12,13. Bones with a tenuous or precarious blood supply
According to Weber, fracture non-unions are cate- (scaphoid, talus) are prone to non-union after
gorised into Hypervascular (hypertrophic) or injury. Sustaining vascularity to fracture ends is
Avascular (atrophic) types.11 The hypervascular group vital. Soft tissue injury from the original traumatic
has adequate vascularity and biological activity to pro- insult and periosteal stripping during surgical fix-
gress to union but limited by bony stability evident on ation impairs this.18 Open fractures disrupt the soft
radiographs with excessive callus in response to motion tissue envelope, where a larger zone of injury is asso-
at the fracture site. Callus is a function of both the mag- ciated with a greater risk of non-union, as per
nitude and frequency of inter-fragmentary motion.14 Gustilo–Anderson classification.18
Inter-fragmentary motion disrupts establish healing, Stability at the fracture site preserves the soft tissue
particularly the empty scaffold within the radiographic envelope and residual vascularity. Mechanical stability
fracture line.2 Weber further subdivided by radiological is required to prevent further insult to callus and

Downloaded from tra.sagepub.com by guest on June 25, 2015

4 Trauma 0(0)

Table 2. The risk factors for fracture non-union.8,15,20 non-union scoring system has no statistical basis, level
4 evidence being constructed by an expert panel. Thus,
Patient factors Fracture-specific factors
there has been no assessment of the conditional rela-
Increasing age High-energy trauma or injury tionship between factors and the magnitude of effect.
severity score There is scope for validation of this tool.
Smoking Soft tissue injury. Larger zone
of injury and high Gustilo-
Anderson Grade
Diagnosis
Poorly diabetic control Large inter-fragmentary gaps Although there is variation in the healing time of frac-
Osteoporosis Biomechanical instability tures according to different skeletal sites, almost all
Vitamin D, calcium or Infection fractures should heal within 3 to 4 months.2 Non-
protein deficiency union is when the fracture has failed to heal by nine
Increased alcohol Prolonged immobilisation months from the time of injury. The timescale in clin-
consumption ical practise is interpreted amongst surgeons, consider-
Reduced muscle mass Perioperative or prolonged ing the patients’ age, location and type of fracture.
and mechanical non-steroidal anti-inflam- Hence, the difficulty reflected in the literature when
stimuli matory drugs (NSAID) use establishing consistency in definitions for non-union.
Post-menopausal Complex of comminuted Diagnosis is subjective and dependant upon the clin-
females fractures icians’ experience. Clinical examination is the corner-
Diaphyseal fractures stone of diagnosis where we assess for tenderness at the
Large fracture haematoma fracture site, movement at the fractures site and pain
whilst applying physiological stresses or whilst weight
baring. Plain radiographs provide additional informa-
modify osteogenic behaviour to promote healing. Perren tion. Assessing for fracture consolidation involves eval-
states that bone cells are sensitive to their environment, uating the extent of bridging callous across the fracture
where mesenchymal cells alter differentiation according site in two orthogonal views and diminishing fracture
to local forces.19 Hydrostatic forces promote a chondro- lines. Radiographs are accessible and cheap but poor at
genic pathway; however, sheer or tensile forces have a identifying non-unions, particularly in the early
detrimental effect resulting in fibrogenesis. Achieving stages.14 High-resolution CT is more sensitive and
mechanical stability in osteoporotic bone is challenging, used in cases where non-union is suspected. DEXA
and technology such as locking screws where the screw scanning is useful to distinguish between hypervasulcar
engages into an osteosynthesis device has increased stiff- and atrophic non-union in the early phase. Diagnostic
ness of fixation. However, in addition to altered compos- adjuvants have been developed (orthometer, strain
ition and architecture, osteoporotic bone has altered gauge bar) with variable implementation clinically.
healing potential.20 Smoking increases the incidence of Attention has been drawn to biochemical markers.
osteoporosis with associated increase in insufficiency The majority of studies are laboratory-based with exten-
fractures. There is a 5–10% bone density deficit in sive work carried out on rat models. Encouraging data
patients who smoked compared with patients who have been reported on the predictive value of transform-
were non-smokers. Nicotine inhibits alkaline phosphat- ing growth factor (TGF), where after an initial peak at
ase (ALP) and collagen production and stimulates two weeks, post-fracture serum levels decrease over four
deoxyribonucleic acid synthesis, possibly secondary to weeks. A rapid decline in serum concentration occurs in
increased cell death and turnover.7 Nicotine prevents delayed union, with significantly lower TGF levels
cellular proliferation during the fracture heating process, detected in non-union groups four weeks after trauma.
altering the maturing of macrophages and fibroblasts Thus, TGF could be used as a predictive cytokine for
and acting directly on osteoblasts. It is also a vasocon- delayed fracture healing.14 In animal models, ALP
strictor agent, causing an alteration of the tissue perfu- returns to normal levels within two months of a fracture,
sion with consequent hypoxia and ischemia.15 but has been found to have a prolonged increase due to
Numerous risk factors have been hypothesised in the delayed union in dogs.14 In clinical trials, ALP showed a
literature to date. However, there is a paucity of evi- significantly higher baseline and a tendency to remain
dence to accurately predict impending non-unions. high in delayed union patients.14
Calori devised a scoring system to facilitate homogen-
eity in assessment of patients and research definitions.21
Current concepts in treatment
The predictive tool combines many of the features listed
in Table 2 to facilitate the assessment and treatment The management of fracture non-union is complex and
planning in patients at risk of non-union. This particular expensive. The trend has shifted towards treatment in

Downloaded from tra.sagepub.com by guest on June 25, 2015

Nandra et al. 5

tertiary centres with additional resource and expertise. In into reamed intramedullary nailing with or without
additional to AO principals of achieving anatomical rhBMP-2 and found no difference in fracture healing,
reduction, obtaining stabile fixation, preserving blood 60% and 48% at week 13 (p ¼ 0.0541) and 68% and
supply and early mobilisation, the non-union surgeon 67% at week 20.26 Lyon found no additional benefit
has to consider the local environment. There are numer- when adding rhBMP-2 and a new, injectable calcium
ous techniques and devices to offer additional stability phosphate matrix (CPM) formulation in acute closed
and to augment local biology. In most circumstances, tibial diaphyseal fractures treated with reamed intrame-
both approaches are adopted. This article discusses dullary nail fixation.27
three treatment adjuncts, of which one is non-invasive. Research to date has focused on acute tibial frac-
tures or tibial non-unions. The use of BMP in acute
open fractures appears to be favourable,22–25 with
Bone morphogenic protein
reduced need for secondary interventions and treatment
Bone morphogenic protein (BNP) is a protein synthe- failure for the treatment of open tibial fractures treated
sised naturally, stimulating the formation of bone and with IM nailing.
cartilage. The protein is a derivative of the transform- Arrington et al. reported a retrospective review of
ing growth factor beta (TGF-ß) superfamily, secreted 414 consecutive cases of iliac crest bone graft proced-
by cells and serves as a signalling agent that influences ures. Donor site morbidity is a concern with 10%
cell division, matrix synthesis and tissue differenti- minor superficial infections, superficial seromas and
ation.22–24 There are two main recombinant BMP prod- minor hematomas and 5.8% major herniation of
ucts in use clinically, BMP-7 and rhBMP-2. These abdominal contents, vascular injuries, deep infections,
induce bone through two pathways. The protein or haematoma or neurologic injuries in their study
induces recruitment of mesenchymal cells from sur- population.28 When considering donor site morbidity
rounding muscle, bone marrow or blood vessels and and scenarios where donor graft is limited, BMP is a
either differentiates these cells into osteoblasts to viable, safe alternative or adjuvant with similar out-
make bone directly or via cartilage cells which subse- comes to autograft.23,29
quently mineralises into bone cells. BMPs also help in
bone matrix production and vascularisation.24
Platelet-rich plasma
Garrison published a review in 2010 comparing BNP
to standard techniques such as iliac crest autograft. The The interest in using platelet-rich plasma (PRP) to aug-
11 studies included randomised controlled trials invol- ment bone healing parallels an increasing understand-
ving patients with acute tibial fractures, tibial non- ing of the role of growth factors in the repair process.
unions, critical size defects and radial mal-unions. The fracture haematoma, containing platelets, is crucial
The authors highlight heterogeneity in methodology, to the initial phase of fracture healing and can often be
surgical treatment, BNP dosage and definition of debrided away during surgical procedures. It does
union. Union rates and fracture healing were compar- therefore seems feasible to replenish the fracture site
able between BNP treatment and control groups. The with pro-inflammatory cells. Platelets have a funda-
pooled relative risk (RR) for achieving union was 1.02 mental role in the release of growth factors. Some of
(95% CI 0.90–1.15). There was some evidence for these such as platelet-derived growth factor (PGDF)
increased healing rates, without requiring a secondary and TGF-ß, vascular endothelial growth factor
procedure, in BMP compared to control groups in (VEGF) and insulin-like growth factor (IGF) are well
acute, open, tibial fractures (RR 1.19, 95% CI 0.99– known.30 These proteins are histo-promotive factors.
1.43).24 This outcome was heavily weighted by one par- They set the stage for tissue healing which includes cel-
ticular study by Govender which suggests that there is a lular chemotaxis, proliferation and differentiation,
dose-dependant effect on outcome, and patients receiv- removal of tissue debris, angiogenesis and the laying
ing 1.50 mg/mL rhBMP-2 group where there was a down of extracellular matrix.30
44% reduction in primary outcome (revision surgery PRP is an autologous blood product, which contains
for delayed or non-union within 12 months of index concentrates of platelets far superior to physiological
surgery), fewer invasive interventions (p ¼ 0.0264) and blood. A concentration of 1,407,640 mL (SD 320 100)
faster fracture-healing (p ¼ 0.0022) than control of platelets in plasma has been suggested to be the
patients.25 However, the cohorts were not standardised, working definition of PRP. This is a platelet count
with a larger proportion of patients treated with five times higher than that of blood, which is normally
reamed intramedullary nails in the 1.5 mg/ml group 150,000 mL to 350,000/mL, with a mean of 200,000/
and mean patient age was younger. Reaming produces mL.30 Harvesting PRP is relatively simple, cost-effective
bone debris which itself acts as an autograft. Aro ran- and safe which has encouraged scientists and clinicians
domised 277 patients with acute open tibial fractures to pursue its use in soft tissue injuries and skeletal

Downloaded from tra.sagepub.com by guest on June 25, 2015

6 Trauma 0(0)

injuries.30,31 However, to date there is very little scien- (1.5 MHz) ultrasound with a 1 kHz repetition module
tific evidence for its use in both of these scenarios. In creating a 200 ms signal with an intensity of 30 mW/cm2
fact in a report in 2008, the National Institute of (cf. ultrasound and Doppler medical imaging intensities
Clinical Excellence does not support its use in tendino- of 20 to 800 mW/cm2). The micro mechanical stimulus
pathies and a more recent Cochrane review by Griffin generated induces nano movement and stimulates
failed to find sufficient evidence to advocate routine use osteoactive cells.
in promoting bone healing.32,33 It is disappointing that It is thought that healing is promoted by stimulating
we have no robust randomised controlled trial or clin- the production of growth factors and proteins which
ical evidence to suggest its efficacy, particularly when increase the removal of old bone, increase the produc-
consulting the findings of pre-clinical trials. Animal stu- tion of new bone and increase the rate at which fibrous
dies have found microscopic and macroscopic advan- matrix at a fracture site is converted to mineralised
tages when using PRP therapy, particularly in bone.38 The growth factors induced have particular
combination with other modalities. In a small study, attributes that promote bone healing, and includes the
Nash reports superior callus formation and mechanical synthesis of proteins involved in angiogenesis (vascular
strength in rabbit tibial osteotomies treated with PDGF endothelial growth factor), osteogenesis (osteocalcin),
compared to controls at four weeks.34 Rai experi- bone mineralisation (alkaline phosphatase), matrix pro-
mented using augmented polycaprolactone-tricalcium duction (type-1 collagen and fibronectin), bone consoli-
phosphate scaffolds (PCL-TCP) with PRP to treat seg- dation (prostaglandin E2) and matrix remodelling
mental bone defects. The treatment group had greater (matrix metalloproteinase).39
vascular invasion at two weeks, compared to controls The evidence from in vitro and animal studies sug-
(PCL-TCP), and a greater proportion of PRP-treated gests that low-intensity pulsed ultrasound (LIPUS) pro-
femurs (83%) achieved bone union as compared to duces significant osteoinductive effects, accelerating the
empty scaffold controls with higher torsional stiffness healing process and improving the bone-bending
(33%).35 Sarker created critical size defects (2.5 cm) in strength.40 Early experimental studies reported a bene-
the tibial diaphysis of sheep and treated them with fit with fracture healing and LIPUS. Heybeli found
either collagen combined with PRP or collagen alone osteotomised femurs in rats to heal with superior
(control). At 12 weeks, neither radiological nor bio- callus assessed radiographically with ultrasound ther-
mechanical or microscopic evaluation identified any apy and a dose-dependent response.41 Azuma found
significant difference in bone healing.36 It is worth greater torsional strength at 25 days in rat femurs in
noting that the platelet concentration only increased treatment groups compared to control following oste-
3.5 fold and that PRP was tested in isolation. otomy.42 Histological analysis identified a correlation
Additional benefit may be seen when augmenting func- with duration of LIPUS and callus formation. Angle
tional biomaterials. varied LIPUS intensities (2, 15 and 30 mW/cm2) with
Calori randomised patients with tibial atrophic non- the belief that the soft tissue between transducer and
unions to either recombinant bone morphogenetic pro- bone dampens signal transmission. They found
tein 7 (rhBMP-7) and PRP during surgical fracture increased cell response in all three stages of bone heal-
revision to stimulate bone healing. 60 patients were ing (cell activation, differentiation and mineralisation).
assigned to each treatment arm; the concentration of After five days with daily treatments of 2, 15 and 30
PRP was 8-fold physiological plasma with allogenic mW/cm2, ALP activity, an early indicator of osteoblast
bone graft used as scaffold. Both clinical and radio- differentiation, increased by 79%, 147%, and 209%,
logical union occurred in 52 (86.7%) cases of the respectively, compared to controls, indicating that
rhBMP-7 group compared to 41 (68.3%) cases of the varying intensities of LIPUS are able to initiate osteo-
PRP group, with a lower median clinical and radio- genic differentiation and mineralisation.43
graphic healing time observed in the rhBMP-7 group The National Institute for Health and Care
(3.5 months vs. 4 months and 8 months vs. 9 months, Excellence (NICE) supports the use of Exogen ultra-
respectively.37 sound bone healing system in both fracture non-unions
(failure to unite nine months after initial fracture with no
evidence of healing for three months) and in delayed
Low-intensity pulsed ultrasound union (no evidence of healing for three months). NICE
Low-intensity pulsed ultrasound (LIPUS) is an external reports clinical evidence with high rates of fracture heal-
device that is applied over the skin at the fracture site, ing and a cost benefit of 1164 per patient with fracture
emitting ultrasonic sound waves. It is increasingly being non-union.38 The device has limitations and is only
used to augment healing, particularly in delayed and considered in patients with adequate fracture reduction
non-union patients. The ExogenTM system (Smith and (inter-fragmentary gap < 10 mm) adequate stability of
Nephew, Memphis, TN, USA) emits low-frequency fracture fixation and compliant patients.39 Exogen

Downloaded from tra.sagepub.com by guest on June 25, 2015

Nandra et al. 7

requires application for 20 min a day, and is more is important. Studies to date suggest accelerated healing
favourable than electromagnetic field therapy in patients after elective procedures, reporting a shorter
(ESTIM), which requires 3 h a day.39 time to healing when using LIPUS. Following distrac-
There is a paucity of robust clinical evidence to sup- tion osteogenesis, the total duration of treatment was
port the use of LIPUS in the treatment of non-union, shorter by 43.6 days.52 Forearm osteotomies had a
with much of the scientific evidence drawing conclu- shorter time to cortical union by 27%.53 The interven-
sions from trials in acute fractures or delayed union. tions have shown benefit in shortening time to clinical
The consolidation rate of fractures following LIPUS and radiological union in osteotomies and elective pro-
intervention ranges from 73 to 86%, with considerable cedures, mirroring that of acute fractures but no effect
heterogeneity in definition and indication.44–48 Ebrahim to the outcome of interest (non-union incidence).54
performed a meta-analysis, investigating efficacy of Current literature reflects considerable variance in
LIPUS (12 trials) and ESTIM (11 trials) against stand- study design, and there is heterogeneity in the definition
ard care in patients with acute fractures and established of non-union and delayed union and a lack of stratifi-
non-unions. The authors found neither LIPUS or cation for variables such as age, sex, smoking status,
ESTIM to improve union rates of acute fractures at fracture location and initial surgical intervention. When
3, 6 or 12 months.49 A prospective randomised control assessing the effect of a device to enhance bone healing
trial did not identify any difference in time to healing, in non-union of fractures, we would preferably use
pain scores or return to work in the treatment group reproducible criteria such as radiological union (brid-
versus placebo in acute clavicle fractures treated ging callous on three corticies), the time to healing, time
conservatively.50 to weight baring pain free and avoidance of further
Schofer randomised patients with tibia fracture surgical procedures. Many of the studies regarding
delayed unions (minimum fracture age four months), exogen and similar devices have inferior subjective out-
to LIPUS treatment (51) or placebo device (50). The comes which predispose to measurement bias.
multi-centre trial reports a mean improvement in bone Furthermore, there has been no direct comparison
mineral density of 34% for LIPUS-treated subjects com- between revision surgery or LIPUS in cases of estab-
pared to placebo (p ¼ 0.002). A mean reduction in bone lished non-union.
gap area also favoured LIPUS treatment (p ¼ 0.014).44
In a large study, Mayr reports a union rate of 86% (817/
Conclusion
951) with mean healing time 152 days. In delayed
unions, the rate was higher, 91% (333/366) with shorter Non-union is a significant problem in healthcare sys-
mean healing time of 129 days. The outcome was influ- tems. With limitations to current data collection, we
enced by smoking and certain medications (NSAID, need to devise robust prospective database to accurately
corticoteroids or calcium blockers).51 estimate the incidence of non-unions. In doing this, we
Rutten analysed the union rate of patients with 6- will gain insight into the magnitude of non-unions and
month old tibial fractures, comparing treatment with allocate resources accordingly through provision of ter-
LIPUS (76% united, 52 of 71 cases, p ¼ 0.0001) to tiary treatment centres and in research planning. By
spontaneous healing rates of tibia non-union according focusing on early accurate diagnosis, through a standar-
to the literature (upper limit ¼ 30%). The mean time to dised pathway we can identify trends and risk factors.
consolidation was 184 days (range, 53–750). When The shift towards standardised definitions and data
stratifying for variables, smokers had a lower healing homogeneity will facilitate the generation of predictive
rate (63%) than nonsmokers (84%), although statistic- models. The models will not only risk stratify patients
ally not significant. Similarly, non-unions located in the but highlight a new focus for research. By risk stratifying
middle third of the tibia showed a healing rate of only patients, we can counsel them accordingly and adjust
53% (eight healed and seven not healed) compared with expectations and reduce indirect costs, which may
89% (eight healed and one not healed) for the proxim- reduce the litigation burden. Management plans can
ally located and 77% (36 healed and 11 not healed) for focus on reversible risk factors to enhance outcome of
the distally located non-union fractures. There was also fracture management and patient satisfaction.
no statistical difference in healing rate for evaluated
variables: comminution, open/closed, non-union type,
fracture age, means of fixation at ultrasound treatment Conflict of interest
start, prior surgical procedure interval, patient age, sex None declared.
and body mass index.48 Time to union becomes an
important factor when considering the patient popula- Funding
tion suffering from non-union. For working individuals This research received no specific grant from any funding
or athletes, time to functional activity or return to sport agency in the public, commercial, or not-for-profit sectors.

Downloaded from tra.sagepub.com by guest on June 25, 2015

8 Trauma 0(0)

Provenance and peer review fractures of long bones: retrospective and prospective
Commissioned, internally reviewed. analyses. J Bone Joint Surg Am 1976; 58: 453–458.
19. Perren SM. Evolution of the internal fixation of long bone
fractures. The scientific basis of biological internal fix-
References
ation: choosing a new balance between stability and biol-
1. Rockwood CA, Green DP and Robert W. Local ogy. J Bone Joint Surg Br 2007; 84: 1093–1110.
Complications. In: Bucholz RW, Cour-Brown M (eds) 20. Giannoudis P, Tzioupis C, Almalki T, et al. Fracture
Rockwood and green textbook of fractures in adults. healing in osteoporotic fractures: is it really different? A
Vol. 1, 7th ed. Philadelphia: Lippincott Williams & basic science perspective. Injury Int J Care Injured 2007;
Wilkins, 2009, pp. 586–592. 38: S90–S99.
2. Frölke JP and Patka P. Definition and classification of 21. Calori GM, Phillips M, Jeetle S, et al. Classification of
fracture non-unions. Injury 2007; 38: 19–22. non-union: need for a new scoring system? Injury 2008;
3. USFDA. Guidance document for the preparation of 39: 59–63.
investigationat device exemptions and pre-market appro- 22. Fisher DM, Wong JM, Crowley C, et al. Preclinical and
vat apptications for bone growth stimutator devices. clinical studies on the use of growth factors for bone
Rockvitte: United States Food and Drug repair: a systematic review. Curr Stem Cell Res Ther
Administration, 1988. 2013; 8: 260–268.
4. Hak DJ, Fitzpatrick D, Bishop JA, et al. Delayed union 23. Wei S, Cai X, Huang J, et al. Recombinant human BMP-
and non-unions: Epidemiology, clinical issues, and finan- 2 for the treatment of open tibial fractures. Orthopedics
cial aspects. Injury 2014; 45: S3–S7. 2007; 35: 847–854.
5. Mills LA and Simpson AH. The relative incidence of 24. Garrison KR, Shemilt I, Donell S, et al. Bone morpho-
fracture non-union in the Scottish population (5.17 mil- genetic protein (BMP) for fracture healing in adults.
lion): a 5-year epidemiological study. BMJ Open 2013; 3: Cochrane Database Syst Rev 2010; 16: CD006950.
e002276. 25. Govender S, Csimma C, Genant HK, et al. Recombinant
6. Mills LA and Simpson AH. The risk of non-union per human bone morphogenetic protein-2 for treatment of
fracture in children. J Child Orthop 2013; 7: 317–322. open tibial fractures: a prospective, controlled, rando-
7. Tzioupis C and Giannoudis PV. Prevalence of long-bone mized study of four hundred and fifty patients. J Bone
non-unions. Injury 2007; 38: 3–9. Joint Surg Am 2002; 84: 2123–2134.
8. Perumal V and Roberts C. Factors contributing to non- 26. Aro HT, Govender S, Patel AD, et al. Recombinant
union of fractures. Curr Orthopaed 2007; 21: 258–261. human bone morphogenetic protein-2: a randomized
9. Kanakaris NK and Giannoudis PV. The health eco- trial in open tibial fractures treated with reamed nail fix-
nomics of the treatment of long-bone non-unions. ation. J Bone Joint Surg Am 2001; 4: 801–808.
Injury 2007; 38: 77–84. 27. Lyon T, Scheele W, Bhandari M, et al. Efficacy and
10. Harrison W and Narayan B. Deformity and Non-union safety of recombinant human bone morphogenetic pro-
after long bone fractures: litigation, negligence and costs. tein-2/calcium phosphate matrix for closed tibial diaphy-
J Bone Joint 2014; 96: 3. seal fracture: a double-blind, randomized, controlled
11. Weber BG and Cech O. Pseudarthrosis, pathotogy, bio- phase-II/III trial. J Bone Joint Surg Am 2013; 95:
mechanics, therapy, results. Bern: Hans Huber, 1976. 2088–2096.
12. McKee MD. Aseptic non-union. In: McKee MD (ed.) 28. Arrington ED, Smith WJ, Chambers HG, et al.
AO principtes of fracture management. Stuttgart/ Complications of iliac crest bone graft harvesting.
New York: Thieme, 2000, pp. 749–754. CORR 1996; 329: 300–309.
13. Patey D, Catagni MA, Argnani F, et al. Ilizarov treat- 29. Papanna MC, Al-Hadithy N, Somanchi BV, et al. The
ment of tibial non-unions with bone loss. Clin Orthop use of bone morphogenic protein-7 (OP-1) in the man-
Retat Res 1998; 241: 146–165. agement of resistant non-unions in the upper and lower
14. Zimmermann G, MUller U and Wentzensen A. The value limb. Injury 2012; 43: 1135–1140.
of laboratory and imaging studies in the evaluation of 30. Alsousou J, Thompson M, Hulley P, et al. The biology of
long-bone non-unions. Injury 2007; 38: 33–37. platelet-rich plasma and its application in trauma and
15. Calori GM, Albisetti W, Agus A, et al. Risk factors con- orthopaedic surgery, a review of the literature. JBJS Br
tributing to fracture non-unions. Injury Int J. Care 2009; 91: 987–996.
Injured 2007; 38: 11–18. 31. Everts PA, Brown MC, Hoffmann JJ, et al. Platelet-rich
16. Frangakis EK. Intracapsular fractures of the neck of the plasma preparation using three devices: implications for
femur. Factors influencing non-union and ischaemic platelet activation and platelet growth factor release.
necrosis. J Bone Joint Surg Br 1996; 48: 17–20. Growth Factors 2006; 24: 165–171.
17. Nilsson LT, Johansson A and Strömqvist B. Factors pre- 32. NICE. Interventional procedure overview of autologous
dicting healing complications in femoral neck fractures. blood injection for tendinopathy. Guidance, National
138 patients followed for 2 years. Acta Orthop Scand Institute of Clinical Excellence, NICE 2008.
1993; 64: 175–177. 33. Griffin XL, Wallace D, Parsons N, et al. Platelet rich
18. Gustilo RB and Anderson JT. Prevention of infection in therapies for long bone healing in adults. Cochrane
the treatment of one thousand and twenty-five open Database Systematic Rev 2007; 7.

Downloaded from tra.sagepub.com by guest on June 25, 2015

 Nandra et al. 9

34. Nash TJ, Howlett CR, Martin C, et al. Effect of platelet- 44. Schofer MD, Block JE, Aigner J, et al. Improved healing
derived growth factor on tibial osteotomies in rabbits. response in delayed unions of the tibia with low intensity
Bone 1994; 15: 203–208. pulsed ultrasound: results of a randomized sham-con-
35. Rai B, Oest ME, Dupont KM, et al. Combination of trolled trial. BMC Muskuloskelet Disorder 2010; 11: 229.
platelet-rich plasma with polycaprolactone- tricalcium 45. Nolte PA, Van der Krans A, Patka P, et al. Low-intensity
phosphate scaffolds for segmental bone defect repair. pulsed ultrasound in the treatment of non-unions.
J Biomed Mater Res 2007; 81: 888–899. J Trauma 2001; 51: 693–703.
36. Sarkar MR, Augat P, Shefelbine SJ, et al. Bone forma- 46. Gebauer D, Mayr E, Orthner E, et al. Low-Intensity
tion in a long bone defect model using a platelet-rich pulsed ultrasound: effects on non-unions. Ultrasound
plasma loaded collagen scaffold. Biomaterials 2006; 27: Med Biol 2005; 31: 1391–1402.
1817–1823. 47. Mayr E, Wagner S and Rüter A. Treatment of non-
37. Calori GM, Tagliabue L, Gala L, et al. Application of unions by means of low-intensity ultrasound. Der
rhBMP7 and platelet-rich plasma in the treatment of long Unfallchirurg 1997; 268: 958–962.
bone non-unions: a prospective randomised clinical study 48. Rutten S, Nolte PA, Guit GL, et al. Use of low-intensity
on 120 patients. Injury 2008; 39: 1391–1402. pulsed ultrasound for post-traumatic non-unions of the
38. NICE. EXOGEN ultrasound bone healing system for tibia: a review of patients treated in the Netherlands.
long bone fractures with non-union or delayed healing: J Trauma 2007; 62: 902–908.
assessment report. NICE.org.uk. 2013. http://www.ni- 49. Ebrahim S, Mollon B, Bance S, et al. Low-intensity
ce.org.uk/guidance/mtg12/evidence/exogen-ultrasound- pulsed ultrasonography versus electrical stimulation for
bone-healing-system-for-long-bone-fractures-with-non- fracture healing: a systematic review and network meta-
union-or-delayed-healing-assessment-report2 (accessed analysis. Can J Surg 2014; 57: E105.
21 November 2014). 50. Lubbert PH, van der Rijt RH, Hoorntje LE, et al. Low-
39. Roussignol X, Currey C, Duparc F, et al. Indications and intensity pulsed ultrasound (LIPUS) in fresh clavicle frac-
results for the ExogenTM ultrasound system in the man- tures: a multi-centre double blind randomised controlled
agement of non-union: a 59-case pilot study. Orthop trial. Injury 2008; 39: 1444–1452.
Traumatol Surg Res 2012; 98: 206–213. 51. Mayr E, Frankel V and Rüter A. Ultrasound – an alter-
40. Kumagai K, Ishikawa H, Yamaguchi Y, et al. Low-inten- native healing method for non-unions? Archive
sity pulsed ultrasound accelerates fracture healing by Orthopaedic Trauma Surg 2014; 120: 1–8.
stimulation of recruitment of both local and circulating 52. Dudda M, Hauser J, Muhr G, et al. Low-intensity pulsed
osteogenic progenitors. J Orthopaedic Res 2012; 9: ultrasound as a useful adjuvant during distraction osteo-
1516–1521. genesis: a prospective, randomized controlled trial.
41. Heybeli N, Yeşildağ A, Oyar O, et al. Diagnostic ultra- J Trauma 2011; 71: 1376–1380.
sound treatment increases the bone fracture-healing rate 53. Urita A, Iwasaki N, Kondo M, et al. Effect of low-
in an internally fixed rat femoral osteotomy model. intensity pulsed ultrasound on bone healing at osteotomy
J Ultrasound Med 2002; 21: 1357–1363. sites after forearm bone shortening. J Hand Surg Am
42. Azuma Y, Ito M, Harada Y, et al. Low-intensity pulsed 2013; 38: 498–503.
ultrasound accelerates rat femoral fracture healing by 54. Hannemann PF, Mommers EH, Schots JP, et al. The
acting on the various cellular reactions in the fracture effects of low-intensity pulsed ultrasound and pulsed elec-
callus. J Bone Miner Res 2001; 16: 671–680. tromagnetic fields bone growth stimulation in acute frac-
43. Angle SR, Sena K, Sumner DR, et al. Osteogenic differ- tures: a systematic review and meta-analysis of
entiation of rat bone marrow stromal cells by various randomized controlled trials. Arch Orthop Trauma Surg
intensities of low-intensity pulsed ultrasound. 2014; 134: 1093–1106.
Ultrasonics 2011; 51: 281–288.

Downloaded from tra.sagepub.com by guest on June 25, 2015

View publication stats