Fractures of the pectoral (thoracic) girdle are a common finding in raptors that have experienced a high-velocity frontal impact, commonly with an automobile. The thoracic girdle consists of the scapula, coracoid, and clavicle, although some authors consider the proximal portion of the humerus to be a portion of the thoracic girdle as well. The triosseal canal is formed by the scapula, coracoid, and clavicle, with the tendon of the supracoracoideus muscle running through this canal. The supracoracoideus muscle is vital in raptors because it allows the wing to be lifted during flight.1,2 The pectoral muscles are attached to the clavicle, keel, sternum, and humerus and contract to provide downward motion of the humerus. The coracoid braces the shoulder joint, allowing for efficient flight. Fractures of the thoracic girdle prevent both takeoff and flight, impacting the ability of raptors to hunt and survive in their natural habitat.3
Typical physical examination findings in raptors with fractures of the thoracic girdle include a dropped wing, crepitus at the fracture site, and callus formation (if the injury is chronic). The VD and left lateral–right lateral radiographic views are commonly used to identify fractures of the thoracic girdle. However, in both of these views, the coracoids, clavicles, and scapulas are superimposed on each other and the thorax, resulting in numerous artifactual radiographic lines. Another view described for evaluation of the shoulder joint4,5 in birds is the caudocranial view of the wing, also known as the hanging drop view or leading edge view. In this view, the bird is held in an inverted position with the head oriented toward the floor and the main axis of the body perpendicular to the surface of the x-ray table. The wing is extended and the bird is angled so that the x-ray beam is aligned in a caudocranial direction and centered on the area of the mid-diaphysis of the radius and ulna. In this view, the scapula, coracoid, and humerus can be evaluated with minimal superimposition; however, this view is technically difficult to perform and may result in an image of lesser quality owing to a large field of view.4
In human medicine, an anteroposterior view and at least 1 other view are obtained during radiographic examination of the shoulder joint.6 Multiple oblique views have been described, including transthoracic lateral, 45° posterior oblique, 60° anterior oblique, and apical oblique views,6,7 in an effort to improve imaging of this region. Clinically, we suspect that oblique radiographic views would also be beneficial in imaging the thoracic girdle in raptors, especially when a fracture of the thoracic girdle is suspected. The purpose of the study reported here was to evaluate use of the H view for assessment of raptors suspected to have fractures of the thoracic girdle. Our hypothesis was that the H view in conjunction with a standard VD radiographic view would provide a more accurate evaluation of the thoracic girdle than the standard VD view alone.
Materials and Methods
Twenty-four raptors evaluated at a rehabilitation center because of suspected trauma and inability to fly were included in the study. Species include Eastern screech owl (Megascops asio; n = 3), red-tailed hawk (Buteo jamaicensis; 6), bald eagle (Haliaeetus leucocephalus; 2), turkey vulture (Cathartes aura; 1), black vulture (Coragyps atratus; 5), barred owl (Strix varia; 3), Cooper's hawk (Accipiter cooperi; 3), and sharp-shinned hawk (Accipiter striatus; 1).
For all birds, initial evaluation consisted of a complete physical examination, CBC, and serum biochemical testing. Suspected fractures were temporarily stabilized with a body wrap, figure-8 bandage, or splint, depending on location. Initial treatment consisted of lactated Ringer solution (25 to 50 mL/kg [11.4 to 22.7 mL/lb], SC), meloxicam (0.2 to 0.5 mg/kg [0.09 to 0.23 mg/lb], PO, q 12 to 24 h) once hydrated, and tramadol (5 to 10 mg/kg [2.3 to 4.5 mg/lb], PO, q 12 h) for pain control as needed.8
Once their condition was stable, birds were anesthetized with isoflurane and radiographs were obtained with a direct digital radiographic system.a,b The kilovoltage was adjusted according to patient thickness, but a constant exposure of 1.4 mAs was used. Patients were positioned in dorsal recumbency with the sternum and thoracic vertebrae superimposed. For the VD view, the wings were extended laterally, the legs were extended caudally, and the x-ray beam was oriented perpendicular to the x-ray table. For the H view, the wings and legs were allowed to remain in a resting, flexed position, and the beam was centered at the thoracic inlet and oriented 45° caudally relative to the plane of the x-ray table. The film-object distance was 100 cm for all views (Figure 1).
Photograph illustrating the radiographic technique for obtaining the H view in a raptor suspected to have a fracture of the pectoral (thoracic) girdle. With the bird positioned in dorsal recumbency and the wings and legs in a resting, flexed position, the tube was rotated 45° caudally and the beam was centered on the thoracic inlet. The film-focus distance was 100 cm.
Citation: Journal of the American Veterinary Medical Association 247, 9; 10.2460/javma.247.9.1037
Radiographic images of the thoracic girdle were anonymized by removing the name, patient number, sex, age, species, and date of the examination. Images were then randomized, assigned a unique alphanumeric identifier, and saved in a lossless .jpeg format. Images were then reviewed during 2 sessions by a board-certified radiologist (MdS) blinded to the final diagnosis. During the first session, the VD and H views for each bird were evaluated independently. During the second session, the VD and H views for each bird were evaluated in combination. To minimize memory recall, the 2 sessions were separated by 2 weeks. All radiographs were classified as radiographically normal (Figure 2) or as showing signs of fracture of any bones of the thoracic girdle (Figures 3 and 4). Sensitivity, specificity, positive predictive value, and negative predictive value and their 95% CIs were calculated for each view, with surgical findings or results of postmortem examination considered the gold standard. The Clopper-Pearson (exact) method was used to calculate 95% CIs. The kappa coefficient was calculated to determine agreement between views and the gold standard, with kappa values > 0.75 considered excellent agreement, kappa values between 0.4 and 0.75 considered fair-to-good agreement, and kappa values < 0.4 considered poor agreement.9 In addition, the McNemar test was used to compare the sensitivity of the VD view, H view, and combination of the 2 views (signs of fracture vs no signs of fracture) with the gold standard (ie, fracture present vs no fracture present, as determined by means of surgery or postmortem examination). Statistical analyses were performed with standard softwarec; values of P < 0.05 were considered significant.
Ventrodorsal (A) and H (B) views of a skeletally mature bald eagle (Haliaeetus leucocephalus) evaluated because of an inability to fly. Notice that the intact osseous structures that make up the thoracic girdle are superimposed on the VD view but not on the H view. α = Coracoid. β = Clavicle. ϵ = Scapula.
Citation: Journal of the American Veterinary Medical Association 247, 9; 10.2460/javma.247.9.1037
Ventrodorsal (A) and H (B) views of a skeletally mature red-tailed hawk (Buteo jamaicensis) evaluated because of an inability to fly. A fracture of the left coracoid is visible on both views. However, the coracoid fracture is easier to see on the H view, and the degree of displacement of the fracture fragment is more obvious on this view. Also, a fracture of the left clavicle is visible on the H view but not the VD view.
Citation: Journal of the American Veterinary Medical Association 247, 9; 10.2460/javma.247.9.1037
Ventrodorsal (A) and H (B) views of a skeletally mature barred owl (Strix varia) evaluated because of an inability to fly. A fracture of the right coracoid is visible on the H view but not on the VD view.
Citation: Journal of the American Veterinary Medical Association 247, 9; 10.2460/javma.247.9.1037
Results
Of the 24 birds included in the study, 9 were confirmed, at the time of surgery or by means of postmortem examination, to have a fracture of the thoracic girdle and 15 were confirmed by means of postmortem examination to not have a fracture of the thoracic girdle. Eight of the 9 birds with fractures of the thoracic girdle had a single fracture (2 with clavicular fractures, 4 with scapular fractures, and 2 with coracoid fractures); the remaining bird had fractures of both the coracoid and clavicle.
For the 9 birds confirmed to have fractures of the thoracic girdle, fractures were correctly identified in 8 birds when the VD radiographic views were evaluated alone, 7 birds when the H views were evaluated alone, and all 9 birds when the 2 views were evaluated in combination (Table 1). Thus, sensitivity was 89%, 78%, and 100%, respectively, for the VD view, H view, and combination of the 2 views. For the 15 birds confirmed to not have fractures of the thoracic girdle, radiographs were correctly classified in 12 birds when the VD views were evaluated alone, all 15 birds when the H views were evaluated alone, and 14 birds when the 2 views were evaluated in combination. Thus, specificity was 80%, 100%, and 93%, respectively, for the VD view, H view, and combination of the 2 views.
There was excellent agreement between the H view and the gold standard (kappa = 0.814) and between the combination of views and the gold standard (kappa = 0.913), whereas the agreement between the VD view and the gold standard was only fair to good (kappa = 0.660). However, when evaluating the sensitivity of the VD view, H view, and combination of the 2 views, the distribution of diagnoses (ie, signs of fracture vs no signs of fracture) was not significantly different (McNemar test; P = 0.317, 0.157, and 0.317, respectively) from the gold standard (ie, fracture present vs no fracture present, as determined by means of surgery or postmortem examination).
Diagnostic accuracy of the standard VD radiographic view alone, the H view alone, and the combination of the VD and H views in 24 raptors suspected clinically to have a fracture of the pectoral (thoracic) girdle.
| No. of birds | ||||||||
|---|---|---|---|---|---|---|---|---|
| View | FN | FP | TN | Sensitivity (95% CI) | Specificity (95% CI) | PPV (95% CI) | NPV (95% Cl) | |
| VD | 8 | 1 | 3 | 12 | 89 (52–100) | 80 (52–96) | 73 (39–94) | 92 (64–100) |
| H | 7 | 2 | 0 | 15 | 78 (40–97) | 100 (78–100) | 100 (59–100) | 88 (64–99) |
| VDandH | 9 | 0 | 1 | 14 | 100 (66–100) | 93 (68–100) | 90 (56–100) | 100 (77–100) |
FN = False-negative result. FP = False-positive result. NPV = Negative predictive value. PPV = Positive predictive value. TN = True-negative result. TP = True-positive result.
Radiographs were reviewed during 2 sessions by a board-certified radiologist blinded to the final diagnosis. During the first session, the VD and H views for each bird were evaluated independently. During the second session, the VD and H views for each bird were evaluated in combination.
Discussion
Results of the present study suggested that the H view alone or in combination with the VD view may be useful for radiographic assessment of raptors suspected to have fractures of the thoracic girdle. This was illustrated by the excellent agreement found for both the H view and the combination of views, compared with the gold standard (ie, fracture present or absent, as determined during surgery or postmortem examination).
A radiographic projection similar to the H view was first described in human medicine by Garth et al10 and has been extensively studied since then.6,11,12 Originally called an apical oblique projection, this view allowed detection of 81% of shoulder injuries in 1 study,6 with the injury apparent only on this view in approximately 10% of the cases. This view has also been used in pediatric medicine to evaluate neonatal trauma and has proven to be valuable for detecting nondisplaced clavicular fractures.12–14
With the H view, tilting the radiographic tube 45° avoids superimposition of the scapula, clavicle, and coracoid. Alternatively, raising the caudal aspect of the bird could provide a similar projection if the radiographic equipment is static. However, in our experience, raising the patient resulted in more variable results and the need to repeat the radiographic exposure multiple times.
At the Auburn University and University of Tennessee colleges of veterinary medicine, the H view is obtained in birds suspected to have trauma to the thoracic girdle and is usually acquired at the same time as the VD view. Subjectively, the H view appears to be especially useful when trauma to the coracoid or its attachment is suspected. In our hands, this radiographic view is both technically easy to perform and reliable when used alone or in conjunction with the VD view. As illustrated by the wide variety of body size and conformation for patients included in the present study, this view will provide diagnostic information for a wide variety of avian patients. However, additional studies on nonraptor patients would be useful.
Although we did not detect statistically significant differences in sensitivity between the H view, VD view, and combination of views in the present study, 1 bird in the present study had a fracture of the thoracic girdle that was not detected on the VD view but was visible when the VD and H views were evaluated in combination. Also, 3 birds had false-positive results when the VD view was evaluated alone (ie, fractures were identified radiographically but were not seen at postmortem examination), but in 2 of these birds, the diagnosis was changed to radiographically normal when the VD view was evaluated in combination with the H view. The only false-positive result when the 2 views were evaluated in combination involved an Eastern screech owl, and the small size of the patient could have been a factor.
The low number of patients and wide variety of fractures represent important limitations of the present study. The authors believe that the nonstatistically significant results obtained with the McNemar test are likely to represent a type II error and could potentially have been corrected with an increased number of patients included in the study. Additional studies with larger numbers of raptors are needed to determine whether the H view alone or the combination of VD and H views provides better diagnostic accuracy than the VD alone for thoracic girdle fractures in general or for clavicular, coracoid, or scapular fracture in particular. Further studies are needed to determine whether body size or conformation has an effect on diagnostic accuracy of these radiographic views.
ABBREVIATIONS
| CI | Confidence interval |
| H view | Caudoventral-craniodorsal oblique radiographic view made at 45° to the frontal plane (Cd45V-CrD view) |
| VD | Ventrodorsal |
Siemens Ysio, Siemens, Malvern, Pa.
Innovet Select, Innovet, Chicago, Ill.
SAS, version 9.4, SAS Institute, Cary, NC.
References
1. Orosz SE, Ensley PK, Haynes CJ. Anatomy and surgical approaches to the wing. In: Orosz SE, ed. Avian surgical anatomy: thoracic and pelvic limbs. Philadelphia: WB Saunders Co, 1992; 4–6.
2. Proctor NS, Lynch PJ. The pectoral girdle. In: Lynch PJ, ed. Manual of ornithology: avian structure and function. New Haven, Conn: Yale University Press, 1993; 134–136.
3. Orosz SE. Clinical considerations of the thoracic limb. Vet Clin North Am Exot Anim Pract 2002; 5: 31–48.
4. Silverman S, Tell LA. Radiology equipment and positioning techniques. In: Radiology of birds: an atlas of normal anatomy and positioning. St Louis: Saunders Elsevier, 2010; 6.
5. Farrow CS. The wing: radiography and normal radiography anatomy. In: Veterinary diagnostic imaging: birds, exotic pets and wildlife. St Louis: Elsevier, 2009; 65–79.
6. Kornguth PJ, Salazar AM. The apical oblique view of the shoulder: its usefulness in acute trauma. AJR Am J Roentgenol 1987; 149: 113–116.
7. Carver E, Carver B. The shoulder girdle. In: Medical imaging: techniques, reflection and evaluation: St Louis: Churchill Livingstone Elsevier, 2012; 67.
8. Hawkins MG. Birds. In: Carpenter JW, ed. Exotic animal formulary. St Louis: Elsevier, 2013; 256–281.
9. Fleiss JL, Levin B, Paik MC. The measurement of Interrater agreement. In: Statistical methods for rates and proportions. 3rd ed. Hoboken, NJ: John Wiley & Sons, 2003; 604.
10. Garth JR, William P, Slappey CE, et al. Roentgenographic demonstration of instability of the shoulder: the apical oblique projection (a technical note). J Bone Joint Surg Br 1984; 66: 1450–1453.
11. Sloth C, Just SL. The apical oblique radiograph in examination of acute shoulder trauma. Skeletal Radiol 1989; 9: 147–151.
12. Weinberg B, Seife B, Alonso P. The apical oblique view of the clavicle: its usefulness in neonatal and childhood trauma. Skeletal Radiol 1991; 20: 201–203.
13. O'Neil BJ, Hirpara KM, O'Briain DO, et al. Clavicle fractures: a comparison of five classification systems and their relationship to treatment outcomes. Int Orthop 2011; 35: 909–914.
14. Peterson CJ, Redlund-Johnell I. Radiographic joint space in normal acromioclavicular joints. Acta Orthop Scand 1983; 54: 431–433.
