Introduction
Antebrachial growth deformities (ABGD) are the most common limb malformation in dogs,
with 63% caused by premature closure of the distal ulnar physis.[1] This typically results in radial external torsion, distal valgus, and increased
procurvatum. These deformities then often cause secondary elbow joint incongruities
and/or antebrachiocarpal joint laxity.[2] Correction of ABGD in dogs can be performed progressively using a circular external
fixator,[3] or in a single surgical session using rigid internal fixation.[4]
[5] Computer-aided design (CAD) now facilitates preoperative surgical planning on the
basis of 3D bone data extracted from a CT scan,[6] as well as the design of both patient-specific 3D-printed surgical osteotomy guides
(PSG) and patient-specific implants (PSI).[7]
Traumatic carpal hyperextension injury is a commonly encountered injury in dogs. Failure
of the carpal palmar fibrocartilage is usually the result of jumping down from a height
and overloading of the carpal support structures.[8] It typically manifests with palmigrade or lowered stance and carpal pain. Recommended
treatment of carpal hyperextension injury is usually by salvage pancarpal arthrodesis
(PCA).[9] Simultaneous ABGD correction and PCA has so far only been reported in five small
breed dogs with ABGD and severe antebrachiocarpal joint collateral ligament laxity
to address both these problems in a single surgery.[10] This is the first report of simultaneous ABGD correction and PCA in a large breed
dog with chronic ABGD and recent traumatic carpal hyperextension injury using a 3D-printed
custom-designed PSG and a 3D-printed PSI with an excellent long-term outcome.
Case Report
History
A 4-year-old, male neutered, 26 kg, Labrador–Springer crossbreed dog was referred
for management of a 3-week history of severe left thoracic limb lameness, which had
been acute in onset following a fall in a local quarry. The owner reported that the
dog had had an obvious deformity affecting that injured limb since adoption 1 year
previously, but that this had not been associated with lameness or affected the dog's
exercise tolerance (several hours a day on the owner's farm). Following initial presentation
at the referring veterinary surgeon's clinic, NSAID medication and short leash walks
were prescribed, which failed to improve the lameness.
Clinical Examination and Surgical Planning
On presentation, the dog displayed a severe toe-touching lameness on the left thoracic
limb. Clinical examination confirmed marked left ABGD characterized by procurvatum,
carpal valgus, and external antebrachial rotation (approximately 60 degrees). Additionally,
there was severe carpal hyperextension ([Fig. 1A, B]). A CT scan of both thoracic limbs was obtained to further assess the deformity
and injury. In addition to the ABGD, the scan revealed a dorsal luxation of the second
carpal bone. The DICOM files were exported to medical image processing software (Osirix,
Pixmeo, SARL; Geneva, Switzerland) and a surface-rendered representation of both thoracic
limbs was created; this was exported as a stereolithography file to computer-aided
design software (Netfabb professional, Netfabb GmbH, Parsberg, Germany, and Geomagic
Freeform, 3D Systems, Rock Hill, California, United States), allowing 3D virtual models
of the imaged bones to be created ([Fig. 2A, B]). These models were used to plan concurrent radial deformity correction and osteotomised
pancarpal arthrodesis (PCA). The radial deformity was multiapical, with mid-diaphyseal
frontal and sagittal plane CORAs at similar levels, and an additional frontal plane
CORA at the level of the distal metaphysis. This CORA was sufficiently distal to permit
correction using an oblique distal radial articular surface osteotomy (i.e., nonparallel
to the joint surface in the frontal plane) in combination with an orthogonal proximal
radiocarpal and ulnar carpal bone osteotomy (i.e., perpendicular to the frontal plane).
This osteotomy was planned to also achieve a 10-degree hyperextension arthrodesis
angle, as well as optimized dorsal plane alignment of the manus. A radial mid-diaphyseal
cuneiform closing wedge ostectomy was concurrently planned such that the combined
result was optimized with regard to the orientations of both radial segments and the
manus in all three planes ([Fig. 2C, D]).
Fig. 1 Clinical photograph of the patient taken preoperatively: (A) cranial and (B) lateral view.
Fig. 2 CAD representation of the preoperative left antebrachium: (A) lateral and (B) cranial view. CAD plan of the surgical correction of the left antebrachial deformity
and pancarpal arthrodesis: (C) lateral and (D) cranial view.
Two PSGs were designed. The radial PSG was designed with a footprint to fit in a unique
position onto the cranial aspect of the bone. It also included five 2.4 mm guide channels,
allowing intraoperative placement of correspondingly sized Ellis pins to secure the
PSG to the radius. The holes left by these Ellis pins would subsequently become the
pilot holes for five PSI screws, and their position and orientation were planned to
optimize osteosynthesis strength. The radial PSG had three osteotomy guide planes,
two for the mid-diaphyseal cuneiform osteotomies and one for the distal articular
surface osteotomy ([Fig. 3A, B]). The second PSG was designed to fit onto the dorso-proximal surface of the radiocarpal
and ulnar carpal bones in a unique position. This PSG had a 1.6 mm guide channel for
an Ellis pin, and an osteotomy guide plane ([Fig. 3C]). Once again, the hole left by the pin was planned as a pilot hole for the radiocarpal
bone (RCB) screw in the PSI. The PSI was designed to extend from the proximal radius
to the distal diaphyses of metacarpals III and IV, and the mid-diaphysis of metacarpal
II. The PSI plate was designed as a 2.4 mm/3.5 mm hybrid, allowing for placement of
one RCB and thirteen metacarpal 2.4 mm cortical screws (in three metacarpal prongs),
and four 3.5 mm cortical screws in each radial segment ([Fig. 4A, B]). The thickness of the plate was based on that of 2.4 mm and 3.5 mm limited compression
plates, with an additional 10% as a safety margin. The contact surface of the plate
exactly matched the contours of the cranial/dorsal radial/RCB/metacarpal cortices,
but was marginally offset by less than 1 mm at the levels of the reduced osteotomies
and over the carpal joints ([Fig. 4A]).
Fig. 3 CAD representation of the patient-specific osteotomy guides: (A) lateral view of radius guide, (B) cranial view of the radius guide, and (C) cranio-lateral view of the radiocarpal bone guide.
Fig. 4 CAD representation of the patient-specific 3D printed titanium alloy plate: (A) caudo-cranial view and (B) medio-lateral oblique view.
PSGs and models of the antebrachium (radius/ulna and manus) were 3D-printed using
Form 3B printers (Formlabs, Somerville, Massachusetts, United States). The bone models
were printed in autoclavable High Temperature resin, and the PSGs in BioMed Amber
resin (Formlabs, Somerville, Massachusetts, United States). This material is autoclavable
and biocompatible (EN ISO 10993–1, 3, 5). The PSGs and models were washed with isopropyl
alcohol and UV-cured according to the manufacturer's instructions. Prior to surgery,
the PSGs and models were steam autoclaved according to the manufacturer's recommendations
(134°C for 20 minutes).
The PSI was 3D-printed at an ISO13485-compliant facility in surgical-grade Ti64 alloy
via direct metal laser sintering at 40 micron layer height (EOS M290 printer using
EOS titanium Ti64 alloy Grade 23; ASTM F136 material standard for surgical implants:
UNS R56401); EOS GmbH, Munich, Germany). The plate was heat-treated according to the
printer manufacturer's instructions (ASTM F3301 compliant), micro-bead blasted, and
the screw holes hand finished. Such a PSI can be produced in about 10 days.
Surgical Procedure
The dog was premedicated using methadone (Comfortan, Eurovet Animal Health B.V., 0.3 mg/kg,
intramuscularly [IM]) and acepromazine (AceSedate, Jurox (UK) Ltd, 0.01 mg/kg IM).
Anesthesia was induced using propofol (PropoFlo, Zoetis, 4 mg/kg intravenously [IV])
and maintained with a mixture of isoflurane (IsoFlo, Zoetis) and oxygen after endotracheal
intubation. A constant rate infusion of ketamine (Anesketin, Eurovet Animal Health
B.V., 10 µg/kg/h IV) provided further analgesia, and cefuroxime (Zinacef, GSK, 20 mg/kg
IV) was administered at induction and 90 minutes later. Paracetamol (B.Braun Melsungen
AG, 10 mg/kg IV, BID) and meloxicam (Metacam, Boehringer Ingelheim, 0.2 mg/kg IV,
SID) were initiated. The anesthetized patient was positioned in dorsal recumbency,
and the limb was prepared aseptically in a hanging limb position.
A routine craniomedial approach to the radius was performed[11] with distal extension over the dorsal aspect of the carpus and metacarpal III. Tenotomy
of the adductor pollicis longus tendon was performed. The entire cephalic vascular
bundle was preserved. The soft tissues around the bony landmarks of the distal radius
were elevated as required to ensure a perfect seating of the footprint of the radial
PSG. Distally, this involved elevation of the proximal joint capsule as well as the
extensor retinaculum and soft tissues in the extensor carpi radialis groove. Comparison
of guide position and fit with the autoclaved 3D-printed antebrachial model facilitated
optimal guide positioning in the patient. Five 2.4 mm Ellis pins were inserted through
the guide channels into the radius. The distal radial ostectomy of the articular surface,
including the ulnar styloid process, was performed first, using an oscillating bone
saw ([Fig. 5]). Then, the two mid-diaphyseal radial osteotomies were performed. Thereafter, a
mid-diaphyseal ulna ostectomy was performed at the level of the radial osteotomy,
using a caudolateral approach. The second osteotomy guide was anchored to the RCB
using a 1.6 mm Ellis pin, and the osteotomies of the proximal aspects of the RCB and
ulnar carpal bone were performed using an oscillating saw. After removal of the PSGs,
the radial segments were reduced and manually aligned with the plate. Cortical screws
of 3.5 mm were placed into the corresponding predrilled pilot holes in the radius,
and a 2.4 mm screw in the RCB, which were not yet fully tightened. Full contact between
the plate and the cortex of each segment was verified before the screws were fully
tightened. The additional three radial screws were placed routinely. The RCB and metacarpals
were aligned with the plate, and two 2.4 mm cortical screws were placed in each metacarpal.
These screws were then removed, and the cartilage in the intercarpal and carpometacarpal
joints was removed using a high-speed burr. A cancellous bone graft was harvested
from the ipsilateral proximal humerus and applied to the arthrodesis sites. The RCB
and metacarpals were once again aligned to the plate; the previously removed 2.4 mm
screws were replaced, and the remaining seven screws were placed routinely. ([Video 1]) Subcutaneous tissues were apposed, and skin closure was routine. Postoperative
radiography revealed good reduction and implant positioning. ([Fig. 6A, B]) A modified Robert–Jones bandage was applied and was replaced each week for a total
of 3 weeks. The dog was discharged on a 1-week course of cephalexin and a 2-week course
of paracetamol and meloxicam. Radiographs were obtained 8 weeks following surgery.
([Fig. 6C, D]) These revealed that all metal implants were stable and in place. The bone healing/arthrodesis
at the different osteotomy sites and joint levels was progressing appropriately, although
ossification was not yet complete. The owner was advised to progressively increase
the short lead exercise over the next 8 weeks, when further follow-up radiographs
were recommended. The owner failed to represent the dog at 4 months postoperatively,
as they felt he was sound.
Video 1 CAD animation of the CT scan-based patient-specific osteotomy surgical guide system
for single-session CORA-based mid-diaphyseal antebrachial correction, osteotomised
pan-carpal arthrodesis, and ulna osteotomy. The antebrachium is stabilized using a
patient-specific 3D-printed titanium alloy implant extending from the proximal radius
to three prongs on the metacarpal bones II, III, and IV.
Fig. 5 Intraoperative cranio-medial view of the left antebrachium (distal to the right)
after application of the radial patient-specific osteotomy guide. The distal radial
osteotomy has been performed, and the radiocarpal bone is retracted using a small
Hohmann retractor.
Fig. 6 Radiographs of the left antebrachium: (A) immediate postoperative medio-lateral projection and (B) cranio-caudal projection; documenting appropriate axial antebrachial alignment following
single-session CORA-based mid-diaphyseal antebrachial correction, osteotomised pan-carpal
arthrodesis, and ulna ostectomy. The antebrachium is stabilized using a patient-specific
3D-printed titanium alloy implant extending from the proximal radius to three prongs
on the metacarpal bones II, III, and IV. Radiographs of the left antebrachium: (C) 2-month postoperative medio-lateral projection and (D) cranio-caudal projection. There is no radiographic evidence of implant loosening.
The bony healing and remodeling at the different levels are progressing well, but
are not complete yet. Radiograph of the left antebrachium: (E) 13-month postoperative medio-lateral projection and (F) cranio-caudal projection; documenting stable implant positioning with no evidence
of implant failure. The radial ostectomy, as well as the antebrachiocarpal and carpometacarpal
arthrodeses, have progressed to full bone healing. The ulnar ostectomy gap has reduced
in size and remodeled, although ulnar bone healing appeared incomplete.
Follow-Up 13 months
The patient returned for a follow-up examination 13 months postoperatively. The owners
reported that his gait and activity level had returned to normal. Clinical examination
revealed a completely stable and pain-free thoracic limb with normal conformation
and no lameness ([Fig. 7A] and [B]). The remainder of the orthopedic examination was normal. Radiographs documented
unchanged implant positioning with no evidence of implant failure. The radial ostectomy,
as well as the antebrachiocarpal and carpometacarpal arthrodeses, had progressed to
full bone healing. The ulnar ostectomy gap had reduced in size and had remodeled,
although ulnar bone healing appeared incomplete. ([Fig. 6E] and [F])
Fig. 7 Clinical photograph of the patient taken 13 months postoperatively (A) cranial view and (B) lateral view; displaying straightened leg conformation and full weight-bearing.
Discussion
Simultaneous ABGD correction and PCA have previously only been reported in a small
series of seven limbs in five dogs weighing less than 10 kg, which initially presented
with concurrent ABGD and secondary chronic, severe, carpal laxity.[10] In that series, 3/7 limbs suffered complications (one major complication eventually
resulting in amputation, and two minor complications). The patient described in this
report differed from these previously described cases with respect to substantially
greater bodyweight (26 kg) and the acute, traumatic nature of the concurrent carpal
hyperextension injury, in addition to chronic ABGD.
A key advantage of the reported technique in this patient is the ability to plan and
optimize the required osteotomies in the virtual CAD-based environment, and then to
transfer this finalized plan to surgery via the use of PSGs and a PSI, acting as a
reduction device. Without the use of PSGs, it would have been extremely challenging
to judge the optimal positions and angles of both the radial and especially the articular
surface osteotomies in all planes. Relative alignment of the distal radius and manus
is easier to judge after a burred arthrodesis. However, in this patient, that approach
would have significantly complicated the radial deformity correction since, due to
the differing levels of key frontal and sagittal plane CORAs, a double-level correction
would then have been necessary to achieve an optimal outcome conformation. Even in
cases where oblique articular surface osteotomies are not required, there may be advantages
to osteotomised rather than burred arthrodesis; in particular, the large, closely
apposed surfaces of cancellous bone with potential for greater stability and more
rapid bone healing. Comparative studies with a larger number of cases would be necessary
to determine if these features do indeed result in better outcomes. The achieved bone
healing in this case does at least demonstrate the clinical efficacy of the technique,
with an excellent outcome also in the long term.[12]
The integration of the PSGs with the PSI allowed the plate to act as the reduction
device for both radial segments and the metacarpals. This not only facilitates accurate
transfer of the CAD-planned alignment to surgery, but also avoids the placement of
reduction pins and guides typically required by non-PSI integrated guide systems.
These are challenging to use for metacarpal alignment due to the lack of optimal guide
placement sites, and, on the radius, add surgical time and require greater exposure
as they must be positioned on a different aspect of the bone compared with the plate.
The presence of the predrilled pilot holes and the accurate contour match between
the contact side of the plate and the cortices allowed rapid and precise reduction
of the radial segments onto the plate as the cortical screws were tightened. With
the relative orientations of the radial segments and plate determined, alignment of
the metacarpals with the distal plate prongs facilitated easy, fast, and accurate
completion of the overall planned alignment of the manus.
A further potential advantage of the PSI design used in this case is the ability to
preplan and optimize screw size, number, position, and (for screws placed into predrilled
pilot holes) screw trajectory. This has the potential to improve osteosynthesis strength
and load distribution when compared to traditional off-the-shelf plates. For example,
off-the-shelf PCA plates applied to either one[13] or two[14]
[15] metacarpals allow placement of five or six screws in positions and orientations
according to the plate design. In contrast, the PSI in this report allowed placement
of 13 metacarpal screws in positions determined by the bone conformation. As well
as reducing the stress on each bone/screw interface, along with the greater surface
area of plate/cortex contact (due to the three distal prongs, and the contour match
between the plate underside and the cortex), load transfer will be distributed over
a much greater surface area, and more evenly, compared with a traditional construct.
This might be expected to reduce the risk of preferential screw loading, toggling,
and sequential screw failure, a classic failure mode of standard nonlocked constructs.[16] It is interesting to note that despite the currently perceived superiority of locking
constructs, there is evidence that their performance under cyclic loading is not necessarily
superior to appropriately applied nonlocked constructs in nonosteoporotic bone.[17] Since the PSI described in this report does not generate interfragmentary compression,
the appropriate bone healing demonstrated in this case must have occurred through
secondary bone healing. This feature suggests that the nonlocked PSI construct was
therefore able to maintain an opportune strain environment for osteogenesis within
the osteotomy gaps. The biomechanical features of this type of implant may differ
from conventional locked and nonlocked constructs, and warrant specific investigation.
The use of CAD-based planning, PSGs, and PSIs has several potential disadvantages.
The necessary level of specific technical knowledge regarding virtual surgical planning
in CAD software, guide and implant design, and their manufacture, requires outsourcing
of these processes for the vast majority of surgeons. This inevitably introduces a
short delay of several days before surgery can be performed, and additional expense.
In our clinic, the addition of performing a CT scan and designing and producing a
PSG-PSI system will add about 25% to the overall fee for the owner. However, understanding
such a complex deformity, accurate planning of the correction from radiographs alone,
and most importantly, surgical execution of the correction would be vastly more complex
without a PSG-PSI system and using off-the-shelf implants. This is in line with previous
studies highlighting improved outcomes and reduced surgical times when using PSG corrections
compared with free-hand corrections.[18]
[19]
[20]
From a surgical perspective, close contact between the cortices, the guides, and the
PSI is necessary. This necessitates soft tissue elevation from the bone in the regions
corresponding to the guide footprints and plate. While the guide footprint contact
area is generally quite small, the contact surface of the plate is larger, and it
seems probable that periosteal blood supply will be compromised beneath the plate,
albeit that no adverse effects were clinically evident in this case. Other potential
disadvantages of 3D-printed patient-specific implants include the current impracticality
of performing finite element analysis or biomechanical testing for every custom plate,
and the possibility of variations in microstructure and surface finish that could
affect mechanical strength. However, the authors speculate that the probability of
rapid and complete healing of an osteotomised arthrodesis over that of a burred arthrodesis
should reduce the risk of construct failure due to screw loosening or mechanical plate
failure.
Considering the excellent clinical long-term outcome in this patient, the authors
feel that several points in the planning and individualized treatment of this patient
might be useful for other clinicians faced with a similar clinical presentation or
considering the use of patient-specific guides and implants. Collating data on such
patients in a multi-center case series appears warranted to assess this further and
to continue to improve the clinical outcome of our patients.
