 INSTRUCTIONAL REVIEW: RESEARCH

In vivo models of bone repair

L. A. Mills, This review is aimed at clinicians appraising preclinical trauma studies and researchers
A. H. R. W. Simpson investigating compromised bone healing or novel treatments for fractures. It categorises the
clinical scenarios of poor healing of fractures and attempts to match them with the
From University of appropriate animal models in the literature.
Edinburgh, We performed an extensive literature search of animal models of long bone fracture
Edinburgh, United repair/nonunion and grouped the resulting studies according to the clinical scenario they
Kingdom were attempting to reflect; we then scrutinised them for their reliability and accuracy in
reproducing that clinical scenario.
Models for normal fracture repair (primary and secondary), delayed union, nonunion
(atrophic and hypertrophic), segmental defects and fractures at risk of impaired healing
were identified. Their accuracy in reflecting the clinical scenario ranged greatly and the
reliability of reproducing the scenario ranged from 100% to 40%.
It is vital to know the limitations and success of each model when considering its
application.

An experimental model for studying bone models used for investigating bone repair in
repair needs to reflect the biomechanics and long bones in animals that have been published
the physiology of the particular clinical sce- in the English language; PubMed, OVID and
nario in humans. However, frequently models Google Scholar search engines were used.
are used that do not meet this criterion. Fresh
critical-size-defect models are employed to rep- General model considerations
resent a nonunion, despite the fact that most Age. Table I lists the average time for cessation of
human nonunions do not have a large defect bone growth and life expectancy in various ani-
and, by definition, are not fresh. mals.1 Studies show that the age of an animal
The clinical scenarios can be considered affects both the quality of bone and time for frac-
under the following headings: 1) normal frac- ture repair2,3: not only is mitosis slower in older
ture repair – direct and indirect healing; animals, but also fewer cells are entering the
 L. A. Mills, FRCS(Tr & Orth),
2) delayed union; 3) established hypertrophic mitotic cycle and significantly fewer osteogenic
Paediatric Orthopaedic Fellow nonunion; 4) established atrophic nonunion – precursor cells are produced per mesenchymal
Royal National Orthopaedic
Hospital, Stanmore, Brockley
stiff or mobile (pseudarthrosis); 5) fractures stem cell.3 Therefore, in any study the age of the
Hill, Middlesex HA7 4LP, UK. with a segmental defect; 6) fractures at risk of animal must be carefully controlled.
 A. H. R. W. Simpson, delayed or nonunion, i.e. high-energy and Gender. Hormonal cycles in the female can
DM(Oxon), FRCS(End & Ed), open fractures, infected fractures and fractures have significant influence on bone repair and
MA(Cantab), Professor of
Orthopaedics and Trauma in compromised patients. turnover. Bone mineral density and endochon-
Edinburgh University, There are many factors relating to the host, dral growth are greatly suppressed during the
Department of Orthopaedics
and Trauma, Chancellors local environment, mechanical construct and reproductive cycle, particularly with the first
Building, Little France, the biological and infective situation that con- litter, and if the mother lactates postpartum the
Edinburgh EH16 4SB, UK.
tribute to delayed bone repair; these need to be deficiency is even greater.4 Rats have an accel-
Correspondence should be sent
to Professor A. H. R. W.
taken into account when selecting a model of erated catch-up period between cycles, but
Simpson; e-mail: impaired healing. This review aims to indicate never reach the same level as nulliparous
hamish.simpson@ed.ac.uk
the issues that should be considered with the females.4,5 Ovariectomised rats, especially
©2012 British Editorial Society application of any model, to highlight the range older ones, have delayed healing of femoral
of Bone and Joint Surgery
doi:10.1302/0301-620X.94B7.
of animal models in the literature for the various fractures and reduced bone mineral density
27370 $2.00 clinical scenarios outlined above, and to suggest (BMD)6 and are therefore used as a model for
J Bone Joint Surg Br
an algorithm for choosing a model for a given osteoporotic fractures. It is important when
2012;94-B:865–74. scenario. Our review is based on analysing using female animals to eliminate these

VOL. 94-B, No. 7, JULY 2012 865

866 L. A. MILLS, A. H. R. W. SIMPSON

Table I. Physeal closure, life expectancy and expected time to fracture union in various species1

Physeal closure Life expectancy Expected time to union of

Animal (tibia) (mths) (mths) simple fractures (wks)
Mouse (Imprinting Control Region) 5 18 to 36 3

Rat (Sprague-Dawley) 11 30 to 48 4 to 6

Rabbit (New Zealand White) 6.8 tibia 84 to 96 6 to 7

5.3 femur

Dog (Greyhound) 7.5 tibia 108 to 168 10 to 13

7.3 femur

Sheep (Suffolk x Dorset) 17 180 10 to 14

variables. Males are more territorial and may require sepa- but are used infrequently for studies of bone healing,16,17
rate cages, making them more expensive to keep. and their size, proportionately short limbs and housing
Choice of animal. The choice of species in orthopaedic requirements can be limiting factors, particularly as an
research is varied.7,8 Martini et al9 analysed 21 500 mam- adult pig can weigh about 150 kg.
mal studies and found that the most common choices were Osteotomy versus other fracture technique. By using a
rats (36%), mice (26%), rabbits (13%), dogs (9%), pri- manual/guillotine/impact device, a fracture is given more
mates (3%), sheep, pigs and cats (2% each). inherent stability from the soft-tissue envelope and from the
There is wide variation in the biochemistry, biomechan- interdigitating bony fragments, whereas osteotomy creates
ics and anatomy of normal bone, and healing processes a cleaner, more controlled break.
between and within species do not necessarily reflect the An osteotomy model should be used with caution when
properties of human bone.8,10,11 Sheep have cancellous and investigating trauma. Park et al18 compared the healing pro-
cortical bone, undergo bone remodelling and have a similar cess in the rabbit tibia between an open osteotomy and a
healing rate, but they also have plexiform bone (akin to closed fracture model, and found that there was a significant
woven bone) and fewer Haversian canals, with differences difference in healing both histologically and biomechanically
in bone composition and fracture stress levels.7,8,12 Dogs between the groups. Additionally, the cellular response to an
have similarities to human bone in composition, remodel- oscillating saw may differ from the response to a burr.
ling and architecture, but have a combination of lamellar Open versus closed fracture technique. A popular closed
and plexiform bone, their remodelling is highly variable model uses a guillotine and stabilisation with an intramed-
and their biomechanical properties differ.7,11,13 Bone in rab- ullary (IM) nail,19 which allows containment of the fracture
bits remodels quickly7 and it has a different microstructure haematoma. The open technique allows direct visualisation
from humans.11 Rats have lamellar bone with good cancel- of the alignment of the bone, and the precise local introduc-
lous but less cortical remodelling, and there are significant tion of compounds, but this theoretically creates an open
differences in composition, density and quality.8 fracture and introduces the risks and variables associated
Mice lack a Haversian canal system,14 but are attractive with a surgical procedure.
owing to their low cost, ease of handling, availability of Fracture stabilisation. Fixation with an IM nail may be used
genetic knockout varieties (breeding in which specific genes in animals of all sizes and allows indirect fracture repair,
within the animal have been deactivated) and the increasing although there is interference at the fracture site. In larger
knowledge of their genetic blueprint, but concern has been animal models the proximal cross-screw can be omitted to
raised about their size, with issues of relevance to the induce instability.
human situation when testing bone substitute scaffolds on Plating is used for direct fracture repair; however, as an
such a small scale.15 The biomechanical testing of bone in open technique it will affect the local haematoma forma-
mice requires highly sensitive equipment. tion and hinder radiological assessment. In addition, there
Rats are useful for both long bone and calvarial models. will be weakness through the screw holes on removal of the
They are hygienic and cheap to house, as several females plate that will compromise the testing of biomechanical
can be kept in one cage. Rabbits have a larger skeleton but stress until the holes have filled in.20-22
are still easily housed; however, there are clear size limita- External fixation, both unilateral and circular, has the
tions when assessing implants compared with dogs or advantages of distance from the fracture site, ease of
sheep.7 Cats are an uncommon choice. Dogs are expensive removal and lack of interference with histological, radio-
and demanding to keep, with additional ethical issues logical or mechanical assessment post mortem; the fracture
regarding their use. Pigs have been shown to be a useful can also be created using a closed technique. Plastic ring
model for investigating the systemic response to trauma, fixators have been used to reduce the weight of the frame

THE JOURNAL OF BONE AND JOINT SURGERY

 IN VIVO MODELS OF BONE REPAIR 867

Table II. Indirect healing

Authors % healing achieved Animal Bone Open/closed (method) Method* Issues to consider
20
Gröngröft et al 100 Mouse Femur Open (Gigli saw Flexible bridging plate Not typical indirect healing
osteotomy) on histology; cost of plate
Cheung et al24 100 Mouse Femur Open (manual Ex-fix Intra-operative haemorrhage
fracture) and death from popliteal
artery damage; pin loosening
Histing et al25 100 Mouse Femur Open Locking plate 20% plate failure plate
dislocation; plate cost;
internal callus; IM healing
Holstein et al26 100 Mouse Femur Closed (3-point Locked femoral nail Axial instability, implant
bending) 0.55 mm displacement
27
Holstein et al 100 Mouse Femur Closed (3-point IM mouse screw Cost of implant
bending) 0.5 mm
Manigrasso and 100 Mouse Femur Closed (3-point 0.25 mm locked IM nail 9% surgical error
O’Connor28 bending)
Bonnarens and 100 Rat Femur Closed (500 0.45 mm IM pin Good reproducibility;
Einhorn19 guillotine) possible bending of IM pin
Reed et al29 100 Rat Tibia Open (burr Ex-fix, 1 mm Polyethylene rings
osteotomy) osteotomy
Bak and 93 Rat Tibia Open (3-point Forceps, 0.8 mm 41% excluded, IM pin bend-
Andreassen30 bending) IM pin ing/failure, fracture at
incorrect level
Pelker and 100 Rat Femur Open (manual 0.9 mm IM pin, 10% complication rate,
Friedlaender31 fracture) 10 mm periosteal variety of reasons
stripping
Waters et al32 71 Rabbit Ulna Open (1 mm saw No fixator, splintage by 19% nonunion rate; extensive
osteotomy) radius callus formation
Hart et al33 100 Dog Tibia Open (saw Unilateral ex-fix,
osteotomy) 6× titanium pins
Goodship and 100 Sheep Tibia Open (osteotomy Ex-fix, 3 mm gap Micromovement induced,
Kenwright34 Gigli saw) +/- micromovement controlled mechanical
environment
Schemitsch et al35 100 Sheep Tibia Open (slap hammer, 7 mm IM nail, proximal High-energy model
3-point bending) and distal locking
* Ex-fix, external fixation; IM, intramedullary

Table III. Direct healing

Authors Union achieved (%) Animal Bone Method* Issues

36
Savaridas et al 100 Rat Tibia 4-hole compression plate; peri- Ensure precise technique with
osteal stripping periosteum and osteotomy
Ashhurst et al37 100 Rabbit Tibia 6-hole stainless steel compression 14% fracture rate through screw holes;
DCP OR plastic plate; open saw direct healing in metal DCP group,
osteotomy indirect with plastic plates
Lewallen et al38 100 Dog Tibia Open osteotomy; compression 44% mortality rate; 34% pin loosening
8-hole plate OR unilateral ex-fix in ex-fix group
Perren et al39 100 Sheep Tibia Saw osteotomy; contour plate
+ lag screw
* DCP, dynamic compression plate; ex-fix, external fixation

but may be chewed through, which can be overcome weight-bearing may ensue, resulting in unreliable models of
by using aluminium three-quarter rings.23 The unilateral normal fracture healing.
fixator has a tendency towards excessive micromovement
and instability in small animal models, which may lead to Animal models for different clinical scenarios
an unpredictable number of hypertrophic nonunions. The model selected needs to reflect the relevant patient group.
Plaster casts can be applied rapidly and are non-invasive, Normal fracture repair. The key feature of this type is that it
but have the disadvantage that they can be chewed or soiled. heals without delay or adjunct. It is often used as a control
Fracture models relying solely on the parallel bone (usu- or to evaluate new agents or interventions. Important issues
ally radius or tibia) for complete stability have been used in are deciding between an open or a closed technique, and
several species. They have the advantage of using no foreign between a fracture or an osteotomy. Table II lists indirect
materials, unobscured radiographs and, in the case of a models of bone healing, differentiating between the forma-
closed fracture model, minimal surgical intervention. How- tion of an open or closed fracture.19,20,24-35 All methods
ever, angular deformity, excessive movement and non- except that of Waters et al32 resulted in good union.

VOL. 94-B, No. 7, JULY 2012

868 L. A. MILLS, A. H. R. W. SIMPSON

Table IV. Established delayed union

Delayed (but eventual)

Method*
Author/s union achieved Animal Bone Points of consideration
Röntgen et al40 Not proven Mouse Femur 0.5 mm osteotomy (Gigli saw); 21-day model only; 40% bone in fracture
flexible external fixator gap but not bridged
Lu et al41 Not proven Mouse Tibia Tibial ex-fix, 3-point bending, Acute ischaemia model, 11% mortality
femoral artery resection and 18% morbidity rate; 21-day model
Aro et al42 Not proven Rat Tibia IM pin 0.5 mm or 0.6 mm +/- locking Both diameters of pin resulted in delayed
pin; 3-point bending closed fracture union, poor positive control; 21%
technique excluded due to fracture displacement
Park et al43 5/6 or 2/5 depending on Rabbit Tibia 4-pin double-bar ex-fix, 13 mm gap; Time to union determined by time and
group 6 mm periosteal stripping;variable number of reoperations. May require
times and number of haematoma some refinement.
excision and washout
Rijal et al44 Not proven Rabbit Radius 10 mm bone excised, gap curetted Results of delayed union control group
at 1 week; no additional splintage poorly recorded
Paterson et al45 Not proven Dog Tibia 15 mm resection, silastic spacer, IM pin Spacer was considered to maintain gap
and plaster cast; silastic removed at rather than prevent union; 36%
8 weeks, model run for further complication rate; no results suggestive
4 weeks of any attempt at healing up to 12 weeks
Schell et al46 4/8 delayed union at nine Sheep Tibia External fixator and osteotomy; 3 mm 6-month model; delayed union only
weeks, 3/8 at six months gap, delay induced by mechanical achieved 50% of time; remainder
instability from reduced stiffness developed nonunion
* ex-fix, external fixation; IM, intramedullary

Table V. Established hypertrophic nonunion (HNU)

Author/s HNU % rate Animal Bone and fixation method* Method Points of consideration
42
Aro et al 40% Rat Fibula; no stabilisation Proprioceptive receptor and HNU possibly due to
sciatic nerve denervation by periosteal stripping not
stripping 8 mm periosteum; denervation, non-weight-
fibula bearing bone, poor HNU rate
fracture with scissors
Hietaniemi et al49 100% Rat Femur; IM nail Open osteotomy; 11 mm 4 of 52 had proximal nail
reaming, 7 mm ‘nail’; 4 mm migration; hypertrophic callus
cauterisation ceases at 15/52 and changes
from a hypertrophic to an
atrophic nonunion
Altner et al50 70% (20% atrophic Dog Ulna; no stabilisation Osteotomy 3 mm to 5 mm Non-weight-bearing bone,
nonunion) bone excision, muscle variable nonunion type
interposition
Heckman et al51 100% at 12 weeks Dog Ulna; fibreglass plaster 3 mm osteotomy, Reoperation; may represent
cast periosteal strip, removal of delayed union
gap tissue at 12 weeks
dos Santos Neto 85% HNU, 15% Dog Radius; no stabilisation 3 mm resection Use of bone wax,
and Volpon52 ‘oligotrophic’ osteotomy, 10 mm no complications noted
periosteal strip, bone wax
interposed
Volpon53 54% HNU, 46% Dog Radius; no stabilisation 5 mm osteotomy, 40 mm 32% complication rate
atrophic nonunion periosteal resection including 5% union rate;
inconsistent nonunion
type created
* IM, intramedullary

Table III provides examples of direct healing by open on many factors, including the fracture technique. Models
reduction and internal fixation, with good results; this tech- of delayed union require a positive control group that
nique is more commonly described in larger animals.36-39 demonstrates that the model does eventually unite.
Established delayed union. Delayed union includes bone Many different methods have been used to recreate this
repair after fracture and osteotomy, where time to union is scenario: instability, reduced vascularity, foreign materi-
prolonged but eventually occurs, with return of structural als, reoperation and distraction osteogenesis (Table IV).
integrity and function. It is a clinical diagnosis and relies These models illustrate lack of healing, but few confirm
on establishing the expected time of healing. This results eventual delayed union, and several have flaws in
in wide inter-observer variation (Table IV).40-46 Bhandari their design.
et al47 questioned 444 orthopaedic surgeons and found a Park et al43 created a reproducible delayed union in the rab-
huge variation in the definitions of delayed union and bit by repeated wound irrigation, which delayed the mean
nonunion of the tibia. The expected time to union of a time to bridging from 6.2 weeks to 7.6 weeks with confirmed
simple fracture in an animal model (Table I) will depend delayed union at 10 weeks. In Choi et al’s48 murine distraction

THE JOURNAL OF BONE AND JOINT SURGERY

 IN VIVO MODELS OF BONE REPAIR 869

osteogenesis model for atrophic nonunion (ANU), one control What is important is how closely they reflect the clinical
group showed consistently delayed union. scenario, what the insult is, and how reproducible and
It remains a challenge to find a clinically relevant, relia- reliable the model is.
ble and reproducible technique that results in delayed but Table VI details some of the models of stiff ANU that
eventual full bone bridging. have been used in the literature (some are modifications of
Established hypertrophic nonunion (HNU). HNU is charac- the work of others): 100% nonunion was achieved in
terised by abundant callus formation, visible radiologically, most models, obtained by a variety of techniques, but all
that does not bridge the fracture (Table V).42,49-53 The gap have certain issues that must be considered.29,43,48,56-66
is not freely mobile, being filled with fibrocartilage. Clini- Several authors have used foreign materials to isolate the
cally, HNU arises as a consequence of excess movement at fracture from the surrounding soft tissues: this does not
the fracture site, achieved by using an IM nail without the mimic the clinical setting but does result in a rate of ANU of
locking screws or by relying on the parallel bone alone for 100%.56 Muscle interposition has been noted to contribute
stabilisation. to human nonunion since the 1800s50 and has been used to
The ability of bone wax (traditionally beeswax with create ANU,59 but does not always provide a consistent
almond oil and salicylic acid) to stem bleeding from bone was result. Others have employed a thermal or a chemical
first described by Horsely in 1892.54 Howard and Kelley55 insult.60,64,65 Reoperation has also been used. Boyan et al66
found that it prevented osteogenesis, induced a thin fibrotic adapted Müller, Schenk and Willenegger’s67 original canine
membrane and depressed the inflammatory response. nonunion model of 1968 but also reoperated, excising the
The studies in Table V show that the ability to achieve repair tissue from the gap. Brownlow and Simpson63 and
consistent hypertrophic nonunion was poor: many had Reed et al29 describe similar models in rabbit and rat, respec-
cases of stiff atrophic rather than hypertrophic nonunion tively, by stripping endosteum and periosteum from around
and nonunions among the positive controls. Heckman et al51 the osteotomy site, as might occur in a high-energy injury.
reported a rate of HNU of 100% simply by creating a Three models of mobile ANU (pseudarthrosis) are also
3 mm osteotomy gap, but the model was only given described in four studies shown in Table VI that use either
12 weeks to unite. Hietaniemi et al49 reported a high rate of movement, distraction or instability.68-71 In 1995, Hiet-
HNU in their model, but by one year the abundant callus aniemi et al49 described a model which they termed a hyper-
formation had become atrophic, and the model has been trophic nonunion; in 199868 they used a similar model
used as a ‘pseudarthrosis’ model in other studies. without cautery that led to non-bridging callus and a 100%
Models of atrophic non-union (ANU). Atrophic nonunion is rate of nonunion, with a gap filled with cartilage.
a well-accepted concept in orthopaedics, but defining it with Segmental/critical-size-defect (CSD) model. The definition
clarity is difficult. Human ANU is broadly defined as when a of CSD is the minimum amount of bone loss that will not
fracture shows no attempt at healing and no progression of heal by bone formation in the lifetime of that animal.72
healing or callus formation after an acceptable period of Hollinger and Kleinschmidt73 defined it as a defect with
time, usually judged radiologically with lack of callus and < 10% bony regeneration. The CSD model was first pro-
rounding-off of the fracture ends. What constitutes an posed in 1934 by Key (Key’s hypothesis),74 who stated that
acceptable period of time is highly inconsistent between segmental bone loss 1.5 times the diaphyseal diameter
orthopaedic surgeons.47 In animal models of nonunion it has would lead to nonunion; Toombs et al75 suggested this to be
been defined as being a fracture that will not heal in the life- an overestimation. Einhorn et al76 found that removing
time of that animal. In many small animal studies 16 weeks 20% (6 mm) was adequate for nonunion in the rat femur.
is accepted as a reasonable period of observation, as it is well Table VII illustrates the species-related variation in size of
beyond the expected timeframe for union. the critical gap.76-87
There are two types of ANU, stiff and mobile, which In CSD the gap created is too wide to be bridged; in a
require different approaches to patient management. The model of nonunion bridging is not achieved because of prob-
stiff ANU shows no attempt at healing radiologically, but lems other than the size of the gap, such as poor vascularity
histologically there is tissue across the fracture site and a or stability. The advantages of the CSD are that it is a repro-
certain amount of mechanical stiffness. The second type ducible, single cause for lack of repair, with no need for
again has no radiological signs of healing, but histologically insults such as foreign body insertion or thermal damage.
there is a mobile cystic cavity that offers no mechanical sta- A long bone CSD of 25 mm to 30 mm in sheep88,89 and
bility. Mobile ANU is less common and more typically of 21 mm to 30 mm in dogs90,91 has been found to be effec-
referred to in animal models as pseudoarthrosis; however, tive, whereas in the rabbit a gap of 15 mm in the radius/
this term is used with great variability and often inter- ulna/tibia is reliable.83,92 There is a paucity of data for the
changeably with ANU, without clarity. We would recom- cat and mouse.
mend that ANUs are described as mobile or stiff, and that The CSD model is commonly used in investigating bone
the term pseudarthrosis be used with caution. regeneration as it is a simple way of developing bony non-
All the models employ an insult to the tissues to estab- union. Recently the model’s primary application has been
lish an ANU without creating a critical size defect (CSD). to test the osseo-inductive and osteoconductive capabilities

VOL. 94-B, No. 7, JULY 2012

870 L. A. MILLS, A. H. R. W. SIMPSON

Table VI. Atrophic nonunion (ANU; stiff and mobile)

Authors % ANU achieved Animal Bone Method* Issues to consider

Stiff ANU
Choi et al48 Variable, 60 Mouse Tibia External fixator without latency period Variable bone bridging seen depending
and varying speeds of distraction on speed of distraction with a few
delayed unions by day 27
Oetgen et al56 100 Mouse Femur Part osteotomy/fracture, retrograde IM 1/3 mortality rate; use of diathermy
pin, partial diathermy
57
Garcia et al 100 Mouse Femur Osteotomy, 1.8 mm gap and 4 mm Less consistent result without stripping
periosteal stripping; IM pin and internal or smaller gap, intramedullary and
clip fixation extramedullary tissue stripping
Dickson et al58 97 Rat Femur Osteotomy, periosteal diathermy, Multiple insults to osteotomy site
endosteal stripping, stabilised with including diathermy
ex-fix
Fujita et al59 Unclear Rat Tibia Closed 3-point bending fracture, Lack of quantitative data, initial HNU
IM nail, muscle interposition followed by ANU
Kokubu et al60 100 Rat Femur IM pin, closed 3-point bending fracture, 15% complication rate; use of diathermy
4 mm cautery around fracture site
29
Reed et al 100 Rat Tibia Ex-fix, 1 mm osteotomy, stripped Chewing of nylon device;
periosteum and curetted canal distance no complications
equal to 1× tibial diameter
Latterman et al61 100 Rabbit Tibia 2 mm osteotomy, 30 mm periosteal strip- Extensive IM and periosteal insult with
and Oni62 ping, IM reaming and nailing, foreign body in situ; repeat operation
silastic tubing round fracture at 4/52
(removed at 4 weeks)
Brownlow and 100 Rabbit Tibia Unilateral ex-fix, osteotomy, 2 mm gap, No complications
Simpson63 stripped periosteum and curetted canal
distance equal to 1× tibial diameter
Park et al43 100 Rabbit Tibia 4-pin double-bar ex-fix, 13 mm gap, Repeat reoperation at day 1 and 2,
6 mm periosteal strip, repeat excision haematoma and irrigation
re-operation to debride
Markel et al64 100 Dog Tibia 5 mm bone resection and gap, 15 mm Use of chemical insult
double application of liquid nitrogen,
unilateral ex-fix
Tiedeman et al65 100 Dog Tibia Ex-fix, 6 mm bone resection, Mixed gender and breed
endosteum and periosteum cauterised
66
Boyan et al 100 Dog Radius 3 mm bone excised, reoperation at 1/52 Repeat operative intervention,
to remove gap tissue, fibreglass splintage mechanical instability for ANU model

Mobile ANU
Hietaniemi et 100 Rat Femur Partial osteotomy, partial manual fracture; Most animals run to 9 weeks, a few for
al68,69 reamed 7 mm, loose unlocked 4 mm or several months; gross mechanical
7 mm IM nail; +/-endosteum cauterised instability used +/- cautery; no complica-
tions cited
Cullinane et al70 66 Rat Femur Ex-fix, 3 mm osteotomy, custom fixator Inconsistent healing in control group;
with interfragmentary bending strain Type II cartilage filled gap; 5-week
and micromovement duration
Harrison et al71 100 Rat Femur Ex-fix, osteotomy, 3 mm distraction Bone ends capped; gap mostly void or
fibrous tissue; 5-week duration
* ex-fix, external fixation; IM, intramedullary; HNU, hypertrophic nonunion

of growth factors and proteins in association with bone ping or excising periosteum from the fracture site, crushing
scaffolds and grafts. or removing muscle, ligating arteries and dividing nerves.
Clinically, a CSD model mimics situations where there Many of the models that reflect periosteal stripping have
has been substantial bone loss, either due to trauma or been reported as models of delayed union or nonunion
through surgery for tumour or infection. However, a CSD (Table VIII).18,43,93-98
does not reflect the circumstances where the pathway to Utvag et al94 studied the effect of muscle injury on frac-
osseous regeneration has been arrested in some way, such as ture healing in the rat tibia and found that muscle loss but
due to instability or metabolic disturbance. not crushing significantly affected healing time. Claes et al98
High-energy, comminuted and open injury models. High- studied various forms of fixation on a three-part ‘fracture’
energy and comminuted injuries are associated with greater (osteotomy) in sheep: the external fixator resulted in the
trauma to soft tissues and higher risks of delayed or nonun- fewest complications, and the compression plate produced
ion. In investigating these situations it is important to have the worst outcome. Richards and Schemitsch97 used a seg-
a model that reflects such soft-tissue and periosteal injury. mental canine model and either muscle flap or skin flap
High-energy injuries can be mimicked in models by strip- cover, with rates of nonunion of 25% and 75%, respec-

THE JOURNAL OF BONE AND JOINT SURGERY

 IN VIVO MODELS OF BONE REPAIR 871

Table VIII. High-energy/comminuted/open fracture models

Authors Animal Bone Model and aim* Model considerations

Schindeler et al93 Mouse Tibia Comparison of distal and midshaft open tibial Fast capacity of murine healing. Technical
fracture. IM nail, open 3-point bending fracture challenge of model. No muscle or skin trauma
Utvag et al94 Rat Tibia Osteotomy and IM pin with ST insult. Fibula N resected to create drop foot. Skin
Comparison of IM muscle crushing with coverage was complete and primary closure.
excision on fracture healing Low-energy method
Claes et al95 Rat Tibia Effect of ST trauma on fracture repair. IM nail Closed injury model. High-energy ST injury but no
fixation, 3-point ST/periosteal stripping
bending +/- ST crushing (impaction device)
Utvag et al96 Rat Femur 3-part segmental fracture; reamed, IM pin +/- Soft tissues and skin kept intact, low-energy
periosteal stripping. Study of periosteal stripping fracture model
on healing and vascularity in segmental
fracture
Park et al43 Rabbit Tibia Osteotomy, 3 mm gap, ex-fix. Effects of repeated Pin site fractures. Gap plus debridement
irrigation and haematoma debridement
Park et al18 Rabbit Tibia Open osteotomy with irrigation vs closed Many variables; open vs closed, osteotomy vs
fracture, ex-fix stabilisation for both 3-point bending, +/- irrigation
Richards and Dog Tibia 2.5 cm devascularised segment, plate fixation, Low energy osteotomy. Controlled, non-traumatic
Schemitsch97 flap coverage. Muscle flap vs skin flap for technique
revascularising segmental bone
Claes et al98 Sheep Tibia Segmental osteotomy with DCP/IM nail/bridge Comminuted fracture model but controlled
plate/ex-fix low-energy osteotomy technique with minimal
ST disruption
* IM, intramedullary; ST, soft-tissue; ex-fix, external fixation; DCP, dynamic compression plate

Table VII. Segmental/critical size defect

Defect size; stabilisation

Authors Union rate Animal Bone technique* Considerations
Drosse et al77 0 Mouse Femur 6 mm; plate or ex-fix
Wingerter et al78 Not given Rat Femur 5 mm bone excised; 5-hole Qualitative study; different histology
plate or IM K-wire depending on fixation type; 4-week
run model
Yasko et al79 0% Rat Femur 5 mm burr defect; polyethylene 9/45 were excluded from study;
plate fixation 10% fixation failure
Einhorn et al76 0 Rat Femur 6 mm (20% of the femur); Easy to construct and low cost PMMA
4-pin unilateral ex-fix ex-fix
80
Ibiwoye et al 0 Rat Fibula 6 mm segment excised; Non-weight-bearing bone
no splintage/fixation
Oakes et al81 0 Rat Femur 8 mm defect by burr; Athymic model; 16-week run model
polyethylene plate
82
Ma et al 16% Rabbit Tibia 14 mm defect “created”; 1/6 united; 8-week run model
unilateral fixator
Cook et al83 0 Rabbit Ulna 15 mm segment excised; Non-weight-bearing bone; 12-week
no splinting/fixation run model
Bolander and 0 Rabbit Ulna 20 mm; no splinting/ fixation Proximal callus fused to radius;
Balian84 non-weight-bearing bone
Johnson et al85 11% Dog Radius 20 mm segment excised; 1/9 healed early, others no union at
ex-fix stabilisation 24 weeks
86
Pluhar et al 0 Sheep Tibia 50 mm; IM nail, locked Partial bone ingrowth but resolved
later by complete periosteal excision
Rozen et al87 50% nil, 50% Sheep Tibia 32 mm segment excised Plate fixation; small quantity of bone
minimal found within the gap
* ex-fix, external fixation; IM, intramedullary; PMMA, polymethylmethacrylate

tively. These studies reproduce the endpoint well, but the Bone repair with infection models. Models of bone infection
models do not always reflect the high-energy transfer asso- have been reviewed99,100: there are many models of
ciated with such an injury. osteomyelitis and septic arthritis, but relatively few incor-
Park et al18 compared an osteotomy technique to a porate fracture repair in the presence of infection. Essentially,
closed fracture model in the rabbit and found delayed models for early infection in the presence of trauma to bone
healing, with smaller haematomas and greater periosteal are very similar to the simple fracture models (Table IX).101-105
damage. They also reported that repeated irrigation and Few studies focus on late infection during fracture repair.
debridement led to delayed healing.18,43 It is clear that Compromised host models. Multiple host and clinical fac-
damage and interference to the periosteal and muscle tors are known to impair fracture healing, including diabe-
envelope will have measurable effects on the degree of cal- tes, hypothyroidism, malnutrition, alcohol, smoking and
lus formation, vascularity and inflammatory cascade of drugs such as non-steroidal anti-inflammatory drugs
that model. (NSAIDs). For each of these situations animal models

VOL. 94-B, No. 7, JULY 2012

872 L. A. MILLS, A. H. R. W. SIMPSON

Table IX. Bone repair with infection

Author/s Animal Bone Method*

Chen et al101,102 Rat Femur 6 mm defect, ex-fix, Staphylococcus
aureus
Andriole et al103 Guinea pig Tibia Closed fracture, IM nail, Staph.
aureus inoculum
Worlock et al104 Rabbit Tibia Open fracture, IM nail, Staph. aureus
into fracture site
Southwood et al105 Rabbit Femur 10 mm defect, plate fixation, Staph.
aureus; issues with mortality and
plate bending
* ex-fix, external fixation; IM, intramedullary

Large Brownlow and Simpson63

animal
Stiff
Reed et al29
Small
Atrophic animal
(Table VI) Garcia et al57
Mobile Large
animal
Small
Nonunion Hypertrophic animal Harrison et al71
(Table V)
Large
animal Heckman et al51
Impaired
healing Small
Hietaniemi et al49
animal
Large
Delayed Park et al43
animal
union
(Table IV) Small
animal

Large Pluhar et al86

Segmental animal
defect
(Table VII) Small
Einhorn et al76
Patient animal
scenario
Perren et al39
Large
animal
Direct Ashhurst et al37
(Table III)
Small
animal Savaridas et al36
Normal
healing Schemitsch et al35
Large
animal
Indirect Goodship and Kenwright34
(Table II)
Small
Bonnarens and Einhorn19
animal

Fig. 1

Flow chart showing suitable animal models for each clinical scenario.

of repair have been described and reviewed by Gaston out are available for diseases such as diabetes, for example,
and Simpson.106 as are ones with deficiencies in the immune system (nude
For studies evaluating the effect of the host genotype on mice/rats, severe combined immunodeficiency (SCID) mice)
fracture repair, strains of mice in which specific genes are or animals that enable certain cells to be tracked (green
suppressed are valuable. Mice with specific genes knocked fluorescent protein (GFP) mice).

THE JOURNAL OF BONE AND JOINT SURGERY

 IN VIVO MODELS OF BONE REPAIR 873

Conclusions 23. Mills L, Noble B, Fenwick S, Simpson H. Assesment of a novel angiogenic factor
in a small animal model of atrophic non-union. J Bone Joint Surg [Br] 2008;90-B(Suppl
In conclusion, a variety of animal models for repair of long II):391.
bone fractures are available to the researcher and can be clas- 24. Cheung KM, Kaluarachi K, Andrew G, et al. An externally fixed femoral fracture
sified according to the range of scenarios that are encoun- model for mice. J Orthop Res 2003;21:685–690.
tered clinically. The success achieved with each model varies, 25. Histing T, Garcia P, Matthys R, et al. An internal locking plate to study intramem-
branous bone healing in a mouse femur fracture model. J Orthop Res 2010;28:397–
and this has implications for power calculations performed 402.
in the design of experimental studies. Figure 1 suggests 26. Holstein JH, Matthys R, Histing T, et al. Development of a stable closed femoral
suitable models of bone repair in animals that could be fracture model in mice. J Surg Res 2009;153:71–75.
27. Holstein JH, Menger MD, Culemann U, Meier C, Pohlemann T. Development of
used to represent different clinical scenarios of human a locking femur nail for mice. J Biomech 2007;40:215–219.
bone healing. 28. Manigrasso MB, O'Connor JP. Characterization of a closed femur fracture model
in mice. J Orthop Trauma 2004;18:687–695.
No benefits in any form have been received or will be received from a commer-
cial party related directly or indirectly to the subject of this article. 29. Reed AA, Joyner CJ, Isefuku S, Brownlow HC, Simpson AH. Vascularity in a
new model of atrophic nonunion. J Bone Joint Surg [Br] 2003;85-B:604–610.
30. Bak B, Andreassen TT. Reduced energy absorption of healed fracture in the rat.
References Acta Orthop Scand 1988;59:548–551.
1. Kilborn SH, Trudel G, Uhthoff H. Review of growth plate closure compared with 31. Pelker RR, Friedlaender GE. The Nicolas Andry Award, 1995. Fracture healing:
age at sexual maturity and lifespan in laboratory animals. Contemp Top Lab Anim Sci radiation induced alterations. Clin Orthop Relat Res 1997;341:267–282.
2002;41:21–26.
32. Waters RV, Gamradt SC, Asnis P, et al. Systemic corticosteroids inhibit bone heal-
2. Lu C, Miclau T, Hu D, et al. Cellular basis for age-related changes in fracture repair. ing in a rabbit ulnar osteotomy model. Acta Orthop Scand 2000;71:316–321.
J Orthop Res 2005;23:1300–1307.
33. Hart MB, Wu JJ, Chao EY, Kelly PJ. External skeletal fixation of canine tibial oste-
3. Bergman RJ, Gazit D, Kahn AJ, et al. Age-related changes in osteogenic stem otomies: compression compared with no compression. J Bone Joint Surg [Am]
cells in mice. J Bone Miner Res 1996;11:568–577. 1985;67-A:598–605.
4. Bowman BM, Miller SC. Skeletal mass, chemistry, and growth during and after 34. Goodship AE, Kenwright J. The influence of induced micromovement upon the
multiple reproductive cycles in the rat. Bone 1999;25:553–559. healing of experimental tibial fractures. J Bone Joint Surg [Br] 1985;67-B:650–655.
5. Naylor KE, Iqbal P, Fledelius C, Fraser RB, Eastell R. The effect of pregnancy on 35. Schemitsch EH, Kowalski MJ, Swiontkowski MF, Harrington RM. Compari-
bone density and bone turnover. J Bone Miner Res 2000;15:129–137. son of the effect of reamed and unreamed locked intramedullary nailing on blood flow
6. Meyer RA Jr, Meyer MH, Tenholder M, et al. Gene expression in older rats with in the callus and strength of union following fracture of the sheep tibia. J Orthop Res
delayed union of femoral fractures. J Bone Joint Surg [Am] 2003;85-A:1243–1254. 1995;13:382–389.
7. Pearce AI, Richards RG, Milz S, Schneider E, Pearce SG. Animal models for 36. Savaridas T, Muir AY, Gaston MS, Noble BS, Simpson AHRW. A murine model
implant biomaterial research in bone: a review. Eur Cell Mater 2007;13:1–10. of internal plate fixation [Poster P02]. Presented at British Orthopaedic Research Soci-
ety (BORS) , Manchester, 2008.
8. Aerssens J, Boonen S, Lowet G, Dequeker J. Interspecies differences in bone
composition, density, and quality: potential implications for in vivo bone research. 37. Ashhurst DE, Hogg J, Perren SM. A method for making reproducible experimental
Endocrinology 1998;139:663–670. fractures of the rabbit tibia. Injury 1982;14:236–242.
9. Martini L, Fini M, Giavaresi G, Giardino R. Sheep model in orthopedic research: 38. Lewallen DG, Chao EY, Kasman RA, Kelly PJ. Comparison of the effects of com-
a literature review. Comp Med 2001;51:292–299. pression plates and external fixators on early bone-healing. J Bone Joint Surg [Am]
1984;66-A:1084–1091.
10. Turner CH, Roeder RK, Wieczorek A, et al. Variability in skeletal mass, structure,
and biomechanical properties among inbred strains of rats. J Bone Miner Res 39. Perren SM, Allgower M, Cordey J, Russenberger M. Developments of compres-
2001;16:1532–1539. sion plate techniques for internal fixation of fractures. Prog Surg 1973;12:152–179.
11. Wang X, Mabrey JD, Agrawal CM. An interspecies comparison of bone fracture 40. Röntgen V, Blakytny R, Matthys R, et al. Fracture healing in mice under controlled
properties. Biomed Mater Eng 1998;8:1–9. rigid and flexible conditions using an adjustable external fixator. J Orthop Res
2010;28:1456–1462.
12. Newman E, Turner AS, Wark JD. The potential of sheep for the study of osteope-
nia: current status and comparison with other animal models. Bone 41. Lu C, Miclau T, Hu D, Marcucio RS. Ischemia leads to delayed union during frac-
1995;16(Suppl):277–284. ture healing: a mouse model. J Orthop Res 2007;25:51–61.
13. Kuhn JL, Goldstein SA, Ciarelli MJ, Matthews LS. The limitations of canine tra- 42. Aro H, Eerola E, Aho AJ. Development of nonunions in the rat fibula after removal
becular bone as a model for human: a biomechanical study. J Biomech 1989;22:95– of periosteal neural mechanoreceptors. Clin Orthop Relat Res 1985;199:292–299.
107. 43. Park SH, Silva M, Bahk WJ, McKellop H, Lieberman JR. Effect of repeated irri-
14. Holstein JH, Garcia P, Histing T, et al. Advances in the establishment of defined gation and debridement on fracture healing in an animal model. J Orthop Res
mouse models for the study of fracture healing and bone regeneration. J Orthop 2002;20:1197–1204.
Trauma 2009;23(Suppl):31–38. 44. Rijal KP, Kashimoto O, Sakurai M. Effect of capacitively coupled electric fields on
15. Viateau V, Logeart-Avramoglou D, Guillemin G, Petite H. Animal models for an experimental model of delayed union of fracture. J Orthop Res 1994;12:262–267.
bone tissue engineering purposes. In: Conn P, ed. Sourcebook of models for biome- 45. Paterson DC, Hillier TM, Carter RF, et al. Experimental delayed union of the dog
chanical research. Totowa: Humana Press Inc., 2008:725–735. tibia and its use in assessing the effect of an electrical bone growth stimulator. Clin
16. Tsukamoto T, Pape HC. Animal models for trauma research: what are the options? Orthop Relat Res 1977;128:340–350.
Shock 2009;31:3–10. 46. Schell H, Thompson MS, Bail HJ, et al. Mechanical induction of critically delayed
17. Hauser CJ. Preclinical models of traumatic, hemorrhagic shock. Shock bone healing in sheep: radiological and biomechanical results. J Biomech
2005;24(Suppl):24–32. 2008;41:3066–3072.
18. Park SH, O'Connor K, Sung R, McKellop H, Sarmiento A. Comparison of healing 47. Bhandari M, Guyatt GH, Swiontkowski MF, et al. A lack of consensus in the
process in open osteotomy model and closed fracture model. J Orthop Trauma assessment of fracture healing among orthopaedic surgeons. J Orthop Trauma
1999;13:114–120. 2002;16:562–566.
19. Bonnarens F, Einhorn TA. Production of a standard closed fracture in laboratory 48. Choi P, Ogilvie C, Thompson Z, Miclau T, Helms JA. Cellular and molecular char-
animal bone. J Orthop Res 1984;2:97–101. acterization of a murine non-union model. J Orthop Res 2004;22:1100–1117.
20. Gröngröft I, Heil P, Matthys R, et al. Fixation compliance in a mouse osteotomy 49. Hietaniemi K, Peltonen J, Paavolainen P. An experimental model for non-union
model induces two different processes of bone healing but does not lead to delayed in rats. Injury 1995;26:681–686.
union. J Biomech 2009;42:2089–2096. 50. Altner PC, Grana L, Gordon M. An experimental study on the significance of mus-
21. Remiger AR, Miclau T, Lindsey RW. The torsional strength of bones with residual cle tissue interposition on fracture healing. Clin Orthop Relat Res 1975;111:269–273.
screw holes from plates with unicortical and bicortical purchase. Clin Biomech (Bris- 51. Heckman JD, Ehler W, Brooks BP, et al. Bone morphogenetic protein but not
tol, Avon) 1997;12:71–73. transforming growth factor-beta enhances bone formation in canine diaphyseal non-
22. Johnson BA, Fallat LM. The effect of screw holes on bone strength. J Foot Ankle unions implanted with a biodegradable composite polymer. J Bone Joint Surg [Am]
Surg 1997;36:446–451. 1999;81-A:1717–1729.

VOL. 94-B, No. 7, JULY 2012

874 L. A. MILLS, A. H. R. W. SIMPSON

52. dos Santos Neto FL, Volpon JB. Experimental nonunion in dogs. Clin Orthop Relat 81. Oakes DA, Lee CC, Lieberman JR. An evaluation of human demineralized bone
Res 1984;187:260–271. matrices in a rat femoral defect model. Clin Orthop Relat Res 2003;413:281–290.
53. Volpon JB. Nonunion using a canine model. Arch Orthop Trauma Surg 82. Ma HL, Chen TH, Hung SC. Development of a new method in promoting fracture
1994;113:312–317. healing: multiple cryopreserved bone marrow injections using a rabbit model. Arch
54. Horsley V. Antiseptic wax. BMJ 1892;1:1165. Orthop Trauma Surg 2004;124:448–454.
55. Howard TC, Kelley RR. The effect of bone wax on the healing of experimental rat 83. Cook SD, Baffes GC, Wolfe MW, et al. The effect of recombinant human osteo-
tibial lesions. Clin Orthop Relat Res 1969;63:226–232. genic protein-1 on healing of large segmental bone defects. J Bone Joint Surg [Am]
56. Oetgen ME, Merrell GA, Troiano NW, Horowitz MC, Kacena MA. Development 1994;76-A:827–838.
of a femoral non-union model in the mouse. Injury 2008;39:1119–1126. 84. Bolander ME, Balian G. The use of demineralized bone matrix in the repair of seg-
57. Garcia P, Holstein JH, Maier S, et al. Development of a reliable non-union model mental defects: augmentation with extracted matrix proteins and a comparison with
in mice. J Surg Res 2008;147:84–91. autologous grafts. J Bone Joint Surg [Am] 1986;68-A:1264–1274.
58. Dickson GR, Geddis C, Fazzalari N, Marsh D, Parkinson I. Microcomputed 85. Johnson KD, August A, Sciadini MF, Smith C. Evaluation of ground cortical auto-
tomography imaging in a rat model of delayed union/non-union fracture. J Orthop Res graft as a bone graft material in a new canine bilateral segmental long bone defect
2008;26:729–736. model. J Orthop Trauma 1996;10:28–36.
59. Fujita M, Matsui N, Tsunoda M, Saura R. Establishment of a non-union model 86. Pluhar GE, Turner AS, Pierce AR, Toth CA, Wheeler DL. A comparison of two
using muscle interposition without osteotomy in rats. Kobe J Med Sci 1998;44:217– biomaterial carriers for osteogenic protein-1 (BMP-7) in an ovine critical defect model.
233. J Bone Joint Surg [Br] 2006;88-B:960–966.
60. Kokubu T, Hak DJ, Hazelwood SJ, Reddi AH. Development of an atrophic nonun- 87. Rozen N, Bick T, Bajayo A, et al. Transplanted blood-derived endothelial progeni-
ion model and comparison to a closed healing fracture in rat femur. J Orthop Res tor cells (EPC) enhance bridging of sheep tibia critical size defects. Bone
2003;21:503–510. 2009;45:918–924.
61. Lattermann C, Baltzer AW, Zelle BA, et al. Feasibility of percutaneous gene trans- 88. Donati D, Di Bella C, Lucarelli E, et al. OP-1 application in bone allograft integra-
fer to an atrophic nonunion in a rabbit. Clin Orthop Relat Res 2004;425:237–243. tion: preliminary results in sheep experimental surgery. Injury 2008;39(Suppl):65–72.
62. Oni OO. A non-union model of the rabbit tibial diaphysis. Injury 1995;26:619–622. 89. Schneiders W, Reinstorf A, Biewener A, et al. In vivo effects of modification of
63. Brownlow HC, Simpson AH. Metabolic activity of a new atrophic nonunion model hydroxyapatite/collagen composites with and without chondroitin sulphate on bone
in rabbits. J Orthop Res 2000;18:438–442. remodeling in the sheep tibia. J Orthop Res 2009;27:15–21.
64. Markel MD, Bogdanske JJ, Xiang Z, Klohnen A. Atrophic nonunion can be pre- 90. Brodke D, Pedrozo HA, Kapur TA, et al. Bone grafts prepared with selective cell
dicted with dual energy x-ray absorptiometry in a canine ostectomy model. J Orthop retention technology heal canine segmental defects as effectively as autograft. J
Res 1995;13:869–875. Orthop Res 2006;24:857–866.
65. Tiedeman JJ, Connolly JF, Strates BS, Lippiello L. Treatment of nonunion by per- 91. Lindsey RW, Gugala Z, Milne E, et al. The efficacy of cylindrical titanium mesh
cutaneous injection of bone marrow and demineralized bone matrix: an experimental cage for the reconstruction of a critical-size canine segmental femoral diaphyseal
study in dogs. Clin Orthop Relat Res 1991;268:294–302. defect. J Orthop Res 2006;24:1438–1453.
66. Boyan BD, Caplan AI, Heckman JD, et al. Osteochondral progenitor cells in acute 92. Geiger F, Bertram H, Berger I, et al. Vascular endothelial growth factor gene-acti-
and chronic canine nonunions. J Orthop Res 1999;17:246–255. vated matrix (VEGF165-GAM) enhances osteogenesis and angiogenesis in large seg-
67. Müller J, Schenk R, Willenegger H. Experimental studies on the development of mental bone defects. J Bone Miner Res 2005;20:2028–2035.
reactive pseudarthroses on the canine radius. Helv Chir Acta 1968;35:301–308 (in 93. Schindeler A, Morse A, Harry L, et al. Models of tibial fracture healing in normal
German). and Nf1-deficient mice. J Orthop Res 2008;26:1053–1060.
68. Hietaniemi K. Retarded mineralization cascade in an experimental nonunion: a 94. Utvag SE, Grundnes O, Rindal DB, Reikeras O. Influence of extensive muscle
sequential polyfluorochrome labeling study in rats. Ann Chir Gynaecol 1998;87:236–
injury on fracture healing in rat tibia. J Orthop Trauma 2003;17:430–435.
239.
69. Hietaniemi K, Lehto MU, Paavolainen P. Major fibrillar collagens and fibronectin 95. Claes L, Maurer-Klein N, Henke T, et al. Moderate soft tissue trauma delays new
in an experimental nonunion: an immunohistochemical study. Acta Orthop Scand bone formation only in the early phase of fracture healing. J Orthop Res
1998;69:545–549. 2006;24:1178–1185.
70. Cullinane DM, Fredrick A, Eisenberg SR, et al. Induction of a neoarthrosis by pre- 96. Utvag SE, Grundnes O, Reikeraos O. Effects of periosteal stripping on healing of
cisely controlled motion in an experimental mid-femoral defect. J Orthop Res segmental fractures in rats. J Orthop Trauma 1996;10:279–284.
2002;20:579–586. 97. Richards RR, Schemitsch EH. Effect of muscle flap coverage on bone blood flow
71. Harrison LJ, Cunningham JL, Stromberg L, Goodship AE. Controlled induction of following devascularization of a segment of tibia: an experimental investigation in the
a pseudarthrosis: a study using a rodent model. J Orthop Trauma 2003;17:11–21. dog. J Orthop Res 1989;7:550–558.
72. Schmitz JP, Hollinger JO. The critical size defect as an experimental model for cra- 98. Claes L, Heitemeyer U, Krischak G, Braun H, Hierholzer G. Fixation technique
niomandibulofacial nonunions. Clin Orthop Relat Res 1986;205:299–308. influences osteogenesis of comminuted fractures. Clin Orthop Relat Res
73. Hollinger JO, Kleinschmidt JC. The critical size defect as an experimental model 1999;365:221–229.
to test bone repair materials. J Craniofac Surg 1990;1:60–68. 99. An YH, Kang QK, Arciola CR. Animal models of osteomyelitis. Int J Artif Organs
74. Key JA. The effect of a local calcium depot on osteogenesis and healing of fractures. 2006;29:407–420.
J Bone Joint Surg 1934;16:176–184. 100.Crémieux AC, Carbon C. Experimental models of bone and prosthetic joint infec-
75. Toombs JP, Wallace LJ, Bjorling DE, Rowland GN. Evaluation of Key's hypoth- tions. Clin Infect Dis 1997;25:1295–1302.
esis in the feline tibia: an experimental model for augmented bone healing studies. 101.Chen X, Schmidt AH, Mahjouri S, Polly DW Jr, Lew WD. Union of a chronically
Am J Vet Res 1985;46:513–518. infected internally stabilized segmental defect in the rat femur after debridement and
76. Einhorn TA, Lane JM, Burstein AH, Kopman CR, Vigorita VJ. The healing of seg- application of rhBMP-2 and systemic antibiotic. J Orthop Trauma 2007;21:693–700.
mental bone defects induced by demineralized bone matrix: a radiographic and bio- 102.Chen X, Schmidt AH, Tsukayama DT, Bourgeault CA, Lew WD. Recombinant
mechanical study. J Bone Joint Surg [Am] 1984;66-A:274–279. human osteogenic protein-1 induces bone formation in a chronically infected, inter-
77. Drosse I, Volkmer E, Seitz S, et al. Validation of a femoral critical size defect model nally stabilized segmental defect in the rat femur. J Bone Joint Surg [Am] 2006;88-
for orthotopic evaluation of bone healing: a biomechanical, veterinary and trauma sur- A:1510–1523.
gical perspective. Tissue Eng Part C Methods 2008;14:79–88.
103.Andriole VT, Nagel DA, Southwick WO. A paradigm for human chronic osteomy-
78. Wingerter S, Calvert G, Tucci M, et al. Comparison of two different fixation tech- elitis. J Bone Joint Surg [Am] 1973;55-A:1511–1515.
niques for a segmental defect in a rat femur model. J Invest Surg 2007;20:149–155.
104.Worlock P, Slack R, Harvey L, Mawhinney R. An experimental model of post-
79. Yasko AW, Lane JM, Fellinger EJ, et al. The healing of segmental bone defects, traumatic osteomyelitis in rabbits. Br J Exp Pathol 1988;69:235–244.
induced by recombinant human bone morphogenetic protein (rhBMP-2): a radio-
graphic, histological, and biomechanical study in rats. J Bone Joint Surg [Am] 105.Southwood LL, Frisbie DD, Kawcak CE, et al. Evaluation of Ad-BMP-2 for
1992;74-A:659–670. enhancing fracture healing in an infected defect fracture rabbit model. J Orthop Res
2004;22:66–72.
80. Ibiwoye MO, Powell KA, Grabiner MD, et al. Bone mass is preserved in a critical-
sized osteotomy by low energy pulsed electromagnetic fields as quantitated by in vivo 106.Gaston MS, Simpson AH. Inhibition of fracture healing. J Bone Joint Surg [Br]
micro-computed tomography. J Orthop Res 2004;22:1086–1093. 2007;89-B:1553–1560.

THE JOURNAL OF BONE AND JOINT SURGERY