If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Hand, Elbow, and Microsurgery Unity (icatMA), Institut Català de Traumatología i Medicina de l’Esport (ICATME), Hospital Universitario Dexeus, Barcelona, Spain
Corresponding author: Rita Sapage, MD, Orthopedic and Traumatology Department, Centro Hospitalar de Trás-os-Montes e Alto Douro, Avenida da Noruega, Lordelo, 5000-508, Vila Real, Portugal.
Hand, Elbow, and Microsurgery Unity (icatMA), Institut Català de Traumatología i Medicina de l’Esport (ICATME), Hospital Universitario Dexeus, Barcelona, SpainOrthopedic and Traumatology Department, Centro Hospitalar de Trás-os-Montes e Alto Douro, Vila Real, Portugal
Hand, Elbow, and Microsurgery Unity (icatMA), Institut Català de Traumatología i Medicina de l’Esport (ICATME), Hospital Universitario Dexeus, Barcelona, Spain
Hand, Elbow, and Microsurgery Unity (icatMA), Institut Català de Traumatología i Medicina de l’Esport (ICATME), Hospital Universitario Dexeus, Barcelona, Spain
Hand, Elbow, and Microsurgery Unity (icatMA), Institut Català de Traumatología i Medicina de l’Esport (ICATME), Hospital Universitario Dexeus, Barcelona, Spain
We hypothesized that a vascularized ulnar periosteal pedicled flap (VUPPF) is a versatile graft applicable in adult patients that yields good outcomes and is a reliable alternative to other vascularized bone grafts to reduce both the technical demands and donor site morbidity of other options.
Methods
We reviewed 11 adult patients who underwent surgical treatment of forearm atrophic nonunion with a VUPPF. Patients’ demographics, outcomes (measured by pain on the visual analog scale; Quick Disabilities of the Arm, Shoulder, and Hand score; range of motion; and grip strength), and associated complications were reported.
Results
Of the 11 patients, 5 had previous surgery in an attempt to treat the nonunion with an autologous cancellous bone graft from the iliac crest or olecranon. The average time from nonunion until the VUPPF was 9 months (SD, ±3 months; range, 6–14 months). The mean visual analog scale score improved considerably after surgery (8.7 vs 0.6), and considerable improvement was also noted in the Quick Disabilities of the Arm, Shoulder, and Hand score (50 vs 6). A notable improvement was seen in grip strength after surgery. Pronation/supination also improved considerably between the preoperative assessment and the final postoperative follow-up.
Conclusions
A vascularized ulnar periosteal pedicled flap seems to be a useful and versatile option for a variety of bone union failures of the upper extremity in adults, either at initial presentation or as a salvage technique.
Reconstruction of upper extremity defects can be challenging, particularly owing to the unique anatomical and functional relationship between the forearm bones.
In the forearm, the majority of cases are atrophic in nature and are usually associated with complex injuries (bone loss, poor soft tissue cover, contamination, and infection).
Several surgical options have been described to treat nonunion in the upper extremity, including plate fixation with bone grafting (autologous cancellous or corticocancellous), intramedullary nailing with bone grafting, bone transport, and external fixation.
However, in the presence of an infection, inadequate vascularity of surrounding tissues, or failed prior grafts, vascularized bone grafts (VBGs) may be indicated.
Over the last 10 years, vascularized periosteal flaps (VPFs) have gained special attention in pediatric reconstructive surgery. The angiogenic and osteogenic potential of VPFs to stimulate and accelerate bone union has been documented in animal models and cadaveric studies.
Vascularized periosteal flaps constitute an adaptation of vascularized bone flaps, limited only to the periosteum and sparing the bone. As a result, VPFs may be a simpler technique that might be both reproducible by most surgeons and associated with less donor site morbidity.
High union rates and excellent clinical outcomes have been reported with VPFs in the treatment of complex situations in the pediatric population, such as recalcitrant nonunion, avascular necrosis, or massive bone defects.
The purpose of this study was to assess the outcomes and associated complications of VUPPFs for a series of adult patients with a variety of nonunions of the upper extremity. We hypothesized that the VUPPF is applicable in adult patients, with a good union rate and good clinical outcomes.
Materials and Methods
The protocol of this study was approved by the institutional review board of Hospital Quirón-Dexeus of Barcelona and Hospital Vithas San José of Vitoria-Gasteiz, Spain. Informed consent was signed by all patients. A prospective study was carried out evaluating the outcomes of 11 patients treated with the VUPPF for atrophic nonunions of the upper extremity. Each patient was operated on by 1 of 2 hand surgeons (S.B.-O. or F.S.) at 1 of 2 regional hand surgery institutions (Hospital Quirón-Dexeus of Barcelona, Spain and Hospital Vithas San José of Vitoria-Gasteiz, Spain). The patients were included based on the following criteria: atrophic nonunion <4 cm, nonunion located in the upper extremity from the elbow to the wrist, and age >18 years, with or without a previous unsuccessful procedure for nonunion. Pain persistence and the absence of radiological healing 6 months after the initial injury or a reoperation was defined as nonunion.
Surgical technique
All surgeries were performed using loupe magnification, under brachial plexus block and with tourniquet control, but without the exsanguination of the limb, to maintain the visibility of perforators and small vessels.
Nonunion approach
For the elbow and wrist nonunion, a double approach was used (over the nonunion site, with a separate incision over the dorsal forearm to raise the VUPPF). For diaphyseal nonunion affecting the radius, a single approach was used to raise the flap and to treat the nonunion site. In 6 cases, we removed preexisting hardware and performed a new osteosynthesis. In the remaining cases, because the fixation was stable, we maintained the previous osteosynthesis (Table 1). Still, in 3 of these 5 patients, we removed the screws from the nonunion site in order to resect the fibrous tissue at the nonunion site and to perform the curettage of the sclerotic bone to a bleeding surface. Resected tissue and hardware were sent for culture. In 2 cases, an autologous cancellous graft (from the iliac crest and the distal radius) was also added to the nonunion site. The decision to add autogenous cancellous bone was made intraoperatively based on the size of the bone defect; an isolated VUPPF was used for minimal bone loss, and a VUPPF with autogenous cancellous bone graft was used for defects of >4 cm. In the remaining patients, the nonunion defect was not bone grafted. In all patients, the nonunion was covered with the VUPPF.
Table 1Patients’ Demographics
Patient
1
2
3
4
5
6
7
8
9
10
11
Age, y
21
34
23
51
37
27
19
32
28
48
42
Sex
M
F
M
M
M
F
M
M
F
M
F
Side
R
L
R
R
L
R
R
L
L
R
R
Dominant hand
D
ND
ND
D
ND
D
D
ND
D
D
D
Initial diagnosis
Radius shaft fracture
Open distal humerus fracture
Fourth and fifth carpometacarpal fracture dislocation
Distal humerus fracture
Radius shaft fracture
Olecranon fracture
Open olecranon fracture
Radius shaft fracture
Open olecranon fracture
Radius shaft fracture
Open distal humerus fracture
Initial treatment
ORIF plate
Distal humerus ORIF + olecranon osteotomy and ORIF TBT
CRPP
Distal humerus ORIF + olecranon osteotomy and ORIF TBT
Conservative treatment
ORIF plate
ORIF TBT
ORIF plate + LS
ORIF TBT
ORIF plate
Distal humerus ORIF + olecranon osteotomy and ORIF TBT
Second treatment
Olecranon ORIF plate + iliac crest autograft
Carpometacarpal arthrodesis
ORIF plate + olecranon autograft
OMR LS + olecranon
ORIF plate + iliac crest autograft
Olecranon ORIF plate + iliac crest autograft
Third treatment
OMR + cultures (negative)
Nonunion location
Radius shaft
Proximal ulna
Carpometacarpal joint
Proximal ulna
Radius shaft
Proximal ulna
Proximal ulna
Radius shaft
Proximal ulna
Radius shaft
Proximal ulna
Nonunion classification
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Atrophic
Time to VUPPF, wk
7
13
8
11
6
8
7
14
6
9
12
Osteosynthesis retention
X
X
X
X
X
VUPPF type
Anterograde
Anterograde
Retrograde
Anterograde
Retrograde
Anterograde
Anterograde
Anterograde
Anterograde
Retrograde
Anterograde
Flap size, cm
5 × 2
7 × 2
3 × 3
8 × 2
3 × 2
7 × 2
8 × 2
4 × 2
7 × 2
3 × 3
7 × 2
Pedicle length, cm
4
8
5
9
3
8
10
4
8
4
8
Bone graft/ origin
No
No
Yes/distal radius autograft
No
No
No
No
No
No
No
Yes/iliac crest autograft
Full surgical time, min
72
80
82
102
70
70
92
85
65
105
112
VUPPF harvesting time, min
32
38
28
42
35
30
34
40
33
39
34
Follow-up, mo
12
14
10
15
12
22
10
16
12
20
14
CRPP, closed reduction percutaneous pinning; D, dominant; LS, lag screw; ND, nondominant; OMR, osteosynthesis material removal; ORIF, open reduction internal fixation; TBT, tension band technique.
A longitudinal line between the lateral epicondyle and the distal radioulnar joint was marked with the elbow in flexion, which corresponded to the axis of the flap. A skin incision following this line was made on the posterior forearm. The dissection began at the distal end of the pedicle in the wrist and progressed proximally. At the wrist, the PIV anastomosis with the posterior division of the anterior interosseous vessels was identified (Fig. 1). If the anastomosis is absent, which occurs rarely, the use of retrograde flaps should be aborted, and alternate methods of reconstruction should be used.
In our series, the anastomosis was present in all cases. For diaphyseal defects in the radius, an anterograde VUPPF can still be an option. Posterior interosseous vessels and their branches to the ulna’s periosteum were exposed between the extensor carpi ulnaris and the extensor digiti minimi muscles (Fig. 2). Along its course, cutaneous and muscular branches of PIVs were ligated. The muscular branches of the posterior interosseous nerve were carefully avoided. The dissection was more critical in the proximal third of the forearm because the PIVs lie deeper and in close proximity to the posterior interosseous nerve. At this level, PIVs run on the abductor pollicis longus and are covered by extensor carpi ulnaris and extensor digiti minimi. The VUPPF was made in an anterograde or retrograde manner, depending on whether the nonunion was located proximally (elbow and proximal radius diaphysis) (Fig. 3A ) or distally (wrist and distal radius diaphysis (Fig. 3B), respectively.
Figure 1Posterior interosseous vessel anastomosis with the posterior division of the anterior interosseous vessels (white arrows, pedicle flap). AIV, anterior interosseous vessels; ECA, extensor compartment artery; IC SRA, intercompartmental supraretinacular artery; pdAIV, posterior division of the anterior interosseous vessels.
Figure 2A, B An anterograde VUPPF showing the cambium layer. ∗The removal of an ulna’s periosteum. #Preserved ulna’s periosteum. The white arrows show the pedicle flap. ECU, extensor carpi ulnaris; EDM, extensor digiti minimi.
For the anterograde VUPPF, the pivot point, or the axis of rotation of the flap, is the origin of the PIV, usually from the common interosseous artery, but in approximately 18% of the cases it is from the ulnar artery.
The posterior interosseous flap: a review of 81 clinical cases and 100 anatomical dissections—assessment of its indications in reconstruction of hand defects.
The distance from the pivot point to the proximal edge of the nonunion was measured and corresponded to the length of the pedicle. Once the vascular axis was dissected and the size of the VUPPF was designed according to the nonunion defect, the PIV anastomoses with the posterior division of the anterior interosseous vessels were ligated. A thin muscle sleeve was left to preserve the periosteal vascularity. As distally as possible, a rectangular periosteal flap and PIVs were elevated from proximal to distal in the ulna using a periosteal elevator (Fig. 2).
For the distally based reverse-flow VUPPF, the pivot point of the flap was 2.5 cm proximal to the distal radioulnar joint where the PIVs anastomosed with the posterior division of the anterior interosseous vessels. The distance from this point to the proximal edge of the nonunion was measured and represented the length of the pedicle. Once the vascular axis was dissected and the size of the VUPPF was designed according to the nonunion defect, the PIVs were ligated 2 cm proximal to the axis. A proximal rectangular periosteal flap and PIVs were elevated distally to proximally in the ulna using a periosteal elevator (Fig. 4).
Figure 4Retrograde VUPPF. #Nonosteogenic layer of the VUPPF. ∗Removal of an ulna’s periosteum. The black dashed arrows show the pedicle flap. ECU, extensor carpi ulnaris; EDM, extensor digiti minimi.
The flap size was adapted to each situation. According to the ulna morphology and its distal shape, anterograde flaps were usually longer and narrower, and the retrograde flaps were usually wider and shorter (Table 1). Once completely detached, the flap was rotated and subcutaneously tunneled to the nonunion. The vascular pedicle was inspected to detect any kinking, compression, or tension. The release of the tourniquet allowed an intraoperative assessment to ensure that the blood supply to the periosteal flap was intact. The vascular pedicle was laid ulnarly, whereas the VUPPF was placed transversely to surround the nonunion and was sutured to the adjacent soft tissue (Fig. 5). The subcutaneous layer and the skin were sutured primarily.
Figure 5A VUPPF placed transversely to surround the nonunion. The white dashed arrows show the pedicle flap.
All patients were maintained immobilized with a plaster orthosis (above the elbow for the proximal ulna and thumb spica plaster for all locations) for 2 weeks after the surgery. After that, a removable orthosis was used for 6–7 weeks, allowing for active range of motion (ROM), passive-assisted exercises, and daily activities. At 10 weeks after surgery, unconstrained activities were allowed.
Assessments
All patients were evaluated by an independent senior hand surgeon before surgery, at 2 weeks after surgery, at 3, 6, and 12 months after surgery, and once a year until the final follow-up. Each patient was assessed clinically and radiographically.
Functional objective measurements were recorded before surgery and at 3, 6, and 12 months after surgery for both the affected and contralateral sides. Range of motion was recorded for flexion and extension of the elbow or wrist and for pronation and supination. Range of motion was measured in degrees to the nearest 5°, using a goniometer. A functional evaluation also included a measurement of grip (in kilograms) using standard dynamometers (Jamar Dynamometer, Preston). A correction for hand dominance was applied to right-handed patients, assuming that the dominant hand is 10% stronger than the nondominant hand. For left-handed patients, grip strength was considered equivalent in both hands.
Before surgery and at the 12-month follow-up, an assessment was performed using the Quick Disabilities of the Arm, Shoulder, and Hand (QuickDASH) score (0, no disability; 100, complete disability) and visual analog scale (VAS) rating (0, no pain; 10, severe pain).
All patients had preoperative and postoperative radiographs of the elbow, forearm, or wrist in 2 different views (anteroposterior and lateral) at each follow-up examination. All patients underwent an additional preoperative and postoperative computed tomography scan. Two senior independent musculoskeletal radiologists, blinded to the type of surgical procedure, evaluated all images. Radiological healing was considered to have occurred when the fracture line disappeared with the formation of bony bridging of 3 or 4 cortices. When associated with the absence of pain, the nonunion was assumed to be healed. Intraoperative and postoperative complications were also recorded.
Statistical methods
Categorical variables are presented as frequencies and percentages. Continuous variables are shown as means, SDs, minimums, and maximums.
Results
Patients’ demographics and surgical procedures
Patients’ demographics are described in Table 1. The initial presenting diagnoses were a radius shaft fracture (n = 4; 36.4%), open distal humerus fracture (n = 2; 18.2%), closed distal humerus fracture (n = 1; 9.1%), open olecranon fracture (n = 2; 18.2%), closed olecranon fracture (n = 1; 9.1%), and fourth to fifth carpometacarpal fracture dislocation (n = 1; 9.1%). There were 4 patients who smoked and 2 patients who had type 2 diabetes mellitus. Prior to the VUPPF, patients were treated as described in Table 1. The 3 patients diagnosed with a distal humerus fracture developed a proximal ulna nonunion due to an iatrogenic complication associated with the transolecranon approach used for initial fracture fixation. Nonunion was located in the proximal ulna (n = 6; 54.5%; Fig. 6), radius shaft (n = 4; 36.4%; Fig. 7), and carpometacarpal joint (n = 1; 9.1%; Fig. 8). All cases were atrophic nonunions. Five patients (45.5%) had a prior attempt to treat the nonunion with autologous cancellous bone graft from the iliac crest or olecranon. The average duration between nonunion and VUPPF was 9 months (SD, ±3 months; range, 6–14 months). The VUPPF was performed in anterograde fashion in 8 patients (72.7%) and in retrograde fashion in 3 patients (27.3%). The mean flap size was 11.8 cm2 (SD, ±3.5 cm2; range, 6–16 cm2), with a mean pedicle length of 6.5 cm (SD, ±2.5 cm; range, 3–10 cm). The vascularized periosteal graft was augmented with a bone graft in 2 patients (18.2%), 1 from the iliac crest and 1 from the distal radius). The mean full surgical time was 85 minutes (SD, ±16 minutes; range, 65–112 minutes), with a mean of 34 minutes for VUPPF harvesting (SD, ±6 minutes; range, 24–42 minutes). All microbiological culture results at the index and at the VUPPF revision surgery were negative.
Figure 6Clinical case of proximal ulna nonunion (patient number 11). A Preoperative lateral radiograph. B Preoperative computed tomography scan (white arrow, nonunion). C Two-week postoperative lateral radiograph. D Three-month postoperative lateral radiograph. E Six-month postoperative lateral radiograph. F Six-month postoperative computed tomography scan.
Figure 7Clinical case of distal radius diaphysis nonunion (patient number 5). A and B Preoperative anteroposterior and lateral radiographs. C Preoperative computed tomography scan (white arrow, nonunion). D Two-week postoperative lateral radiograph. E Three-month postoperative lateral radiograph (white arrow, callus formation). F Three-month postoperative computed tomography scan. G Six-month postoperative computed tomography scan.
Figure 8Clinical case of fourth and fifth carpometacarpal nonunion (patient number 3). A and B Preoperative anteroposterior and oblique radiographs (black arrow, nonunion; white arrows, osteolysis around the screw; asterisk, distal radius as origin of bone graft). C and D Preoperative computed tomography scan (white arrows, nonunion). E and F Three-month postoperative anteroposterior and lateral radiographs. G Six-month postoperative computed tomography scan.
Objective measurements are detailed in Table 2. Postoperative values at the final follow-up for grip strength averaged 36 kg (SD, ±8 kg; range, 18–46 kg). An improvement after surgery was seen for grip strength, which was a mean of 18 kg at presentation. The final mean grip strength averaged 99.5% of the contralateral side (SD, ±9%; range, 85%–115%), corrected for hand dominance. Pronation/supination also improved between the preoperative assessment and the final postoperative follow-up. Relative to the preoperative ROM, there was a mean gain of 11° in pronation and 7.5° in supination. At the 12-month follow-up, ROM was similar between the treated and untreated joints.
Patient-reported outcomes are summarized in Table 3. At the 12-month follow-up, postoperative pain on the VAS scale during everyday activities averaged 0.6 (SD, ±0.7; range, 0–2). Mean VAS scores improved after surgery, since the mean VAS at presentation was 8.7 (SD, ±0.9; range, 7–10). Postoperative improvement was also noted in QuickDASH (50 vs 6, compared with presentation). A QuickDASH score ranging from 0 to 29 is thought to be the point at which patients are “no longer considering their upper-limb disorder a problem.”
All nonunions united, as confirmed by computed tomography scan, with 4 (36.4%) healed by 3 months and the remaining 7 (63.6%) healed by 6 months (Table 2).
Postoperative complications were identified in 3 (27.3%) patients: 2 patients developed a keloid scar and 1 patient developed olecranon bursitis. No patient required further surgery.
Discussion
Compression plating and bone grafting (with either cancellous or cortical bone) is an established option for radius and ulna nonunion and is usually associated with excellent outcomes.
For particularly unfavorable scenarios, such as a substantial bone defect, infection, longstanding nonunions, and previously failed nonunion treatment, VBGs have considerably enhanced the union rates compared with those for nonvascularized bone grafts. Hence, a variety of VBG techniques have been used, with published union rates between 91% and 100%.
Demanding microsurgical techniques such as the vascularized fibula bone graft or the vascularized medial femoral condyle graft have also been described.
Local pedicled vascularized flaps offer several advantages over a free flap, including less donor site morbidity, an operative field limited to 1 upper extremity, and a less demanding technique; therefore, they may be easier for most hand surgeons to reproduce.
even in some difficult cases, such as a longstanding nonunion or in patients previously treated for nonunion by another procedure. We assessed patient-reported and objective functional outcomes in addition to radiological evidence of union, which is not consistently done.
Although some reports comment on pain improvement and functional outcomes, it is difficult to compare results because of variations in the measurement of outcomes.
We achieved complete relief of pain and excellent functional outcomes, as measured by the QuickDASH, grip strength and ROM. Although complications after elbow and forearm treatment are rare, some authors report cases of refracture, vascular or nerve injury, infection, and failure to achieve bone union.
In our series, we had minor postoperative complications and none were directly related to the bone union. Carpometacarpal nonunion is rarely reported in the literature.
Even though we describe only a specific case of nonunion following failed fourth and fifth carpometacarpal arthrodesis, we found this to be an encouraging outcome.
Although the PIV axis had initially been described to obtain a VBG, a posterior approach has been developed as a VPF to treat difficult forearm nonunions in children. The rationale for this change is that a VPF conforms more easily to the recipient site configuration than a VBG. The PIV pedicle flap has been successfully used in different ways by changing its flow direction (anterograde, proximally based or retrograde, distally based) and its soft tissue contents (fascia, fasciocutaneous, osteofasciocutaneous).
Improvement of the rotation arch of the posterior interosseous pedicle flap preserving both reverse posterior and anterior interosseous vascular sources.
In our case series, we used a PIV-based periosteal flap to treat bone nonunions, which was adaptable to each case. The flap was anterograde for cases of proximal forearm and elbow and retrograde for cases of the distal forearm and wrist. For that reason, it is important to note the versatility of the VUPPF as a periosteal graft with sufficient length and width to cover small and medium defects from the elbow to the wrist without pedicle tension (as seen by the wide range of 3–10 cm in our case series).
Although different VPFs have been successfully used in especially adverse situations in the pediatric population, their applicability in adults is still unclear.
Composite vascularized fibular epiphyseo-osteo-periosteal transfer for hip reconstruction after proximal femoral tumoral resection in a 4-year-old child.
Vascularized humeral periosteal flap to treat lateral humeral condyle nonunion: an anatomical study and report of two successfully-treated pediatric cases.
The greatest concern relates to the possibly unpredictable bone formation provided by the flap with advancing age, as angiogenic and osteogenic periosteal capacities seem to decrease after epiphyseal closure.
However, in our case series, union was achieved in all cases, with a mean age of 32.9 years (54.5% above 30 years). In these particular scenarios, VPFs appear to have union rates similar to those seen in the pediatric setting.
Composite vascularized fibular epiphyseo-osteo-periosteal transfer for hip reconstruction after proximal femoral tumoral resection in a 4-year-old child.
Vascularized humeral periosteal flap to treat lateral humeral condyle nonunion: an anatomical study and report of two successfully-treated pediatric cases.
A limitation of the present study is the lack of comparison to a control group. Without a control group, we cannot determine whether the nonunions would have healed with cancellous grafting alone. Two patients presented with a bone defect of larger than 4 cm and were simultaneously treated with a corticocancellous bone graft. This can also limit conclusions about VUPPF results as a single procedure, particularly in cases with more extensive bone loss. Our study also analyzed 3 different scenarios of upper limb nonunion together, which also added heterogeneity to the sample.
In conclusion, VUPPF seems to be a useful and versatile technique for a variety of nonunions of the upper extremity in adults, at initial presentation or as a salvage technique, and is associated with low morbidity.
Acknowledgments
The authors thank Dr Fernando Juarez Cesca, from Hospital San Bernardo in Salta, Argentina, for his contribution with the schematic illustration artwork.
References
Faldini C.
Traina F.
Perna F.
Borghi R.
Nanni M.
Chehrassan M.
Surgical treatment of aseptic forearm nonunion with plate and opposite bone graft strut. Autograft or allograft?.
The posterior interosseous flap: a review of 81 clinical cases and 100 anatomical dissections—assessment of its indications in reconstruction of hand defects.
Improvement of the rotation arch of the posterior interosseous pedicle flap preserving both reverse posterior and anterior interosseous vascular sources.
Composite vascularized fibular epiphyseo-osteo-periosteal transfer for hip reconstruction after proximal femoral tumoral resection in a 4-year-old child.
Vascularized humeral periosteal flap to treat lateral humeral condyle nonunion: an anatomical study and report of two successfully-treated pediatric cases.
00167-2&id=gr1.jpg){kind=link}
00167-2&id=gr2.jpg){kind=link}
00167-2&id=gr3.jpg){kind=link}
00167-2&id=gr4.jpg){kind=link}
00167-2&id=gr5.jpg){kind=link}
00167-2&id=gr6.jpg){kind=link}
00167-2&id=gr7.jpg){kind=link}
00167-2&id=gr8.jpg){kind=link}