Radiographic measurement of the cardiac silhouette is a survey diagnostic test that is commonly used in conjunction with physical examination to determine whether a patient likely has cardiac disease. However, physical examination of avians has limitations because of the high heart rate of most species, which makes it difficult to evaluate birds for murmurs or arrhythmias. Radiographs are commonly obtained and evaluated in avian species, but differences among species limit the usefulness of the few reference values currently available.
Radiographic measurements of the cardiac silhouette in avian species were first reported for Amazon parrots in 19941 and have since been determined for other psittacine species (African grey parrots [Psittacus erythacus], Senegal parrots [Poicephalus senegalis], orange-winged Amazon parrots [Amazona amazonica], Spix macaws [Cyanopsitta spixii], and budgerigars [Melopsittacus undulatus]),2–4 select falconry species (common kestrels [Falco tinnunculus], red-tailed hawks [Buteo jamaicensis], Harris hawks [Parabuteo unicinctus], peregrine falcons [Falco cherrug], and lanner falcons [Falco biarmicus])5–8 and other wildlife avian species (screech owls [Otus asio] and Canada geese [Branta canadensis]).7 Ventrodorsal measurements of bald eagles (Haliaeetus leucocephalus) have also been reported.a
Radiographic measurements have been established to provide standards by which to determine cardiomegaly. Common causes of cardiomegaly in wild avian patients include congenital abnormalities (iron storage disease),9 viral disease (avipoxvirus),10 endoparasitism (Plasmodium spp),11 and toxins (fumonisin B1 and moniliformin).12
Ospreys (Pandion haliaetus) are piscivorous raptors in the family Pandionidae of the order Accipitriformes.13 Similar to most of the other species in that order, ospreys are diurnal birds that rely on sight to hunt prey. They have been used as a sentinel species for monitoring aquatic environmental changes since the 1960s because they are apex predators in the piscivorous food chain, have a long life span (the oldest osprey was reportedly 25 years and 2 months old)14 and thus a long-term opportunity for bioaccumulation of lipophilic contaminants, can adapt to nest disturbances and human encroachment, and are found in almost all countries.15 However, the authors are not aware of any reported standards for radiographic measurements of the cardiac silhouette of ospreys. The purpose of the study reported here was to determine reference values for the cardiac silhouette in this species to enable monitoring of diseases and conditions attributable to toxins that may affect ospreys and the ecosystems they inhabit.
Materials and Methods
Animals
Data were collected by retrospective evaluation of records for juvenile and adult ospreys examined at the Clinic for the Rehabilitation of Wildlife in Sanibel, Fla, between January 2015 and March 2018. Standard procedures at arrival included a physical examination, hematologic analysis (manual blood count, PCV, and total solids concentration), examination of feces, and radiographic evaluation.
Birds included in the study had no evidence of cardiac disease as determined on the basis of the history (eg, no known exposure to toxins) and results of a physical examination (no dyspnea, murmurs or arrhythmias, or coelomic distention) and manual blood count (no polycythemia, anemia, or blood parasites). Because of the potential impact of dehydration on size of cardiac chambers, birds with a plasma protein concentration > 4 g/dL were excluded.
Radiography
Radiographs were obtained while birds were anesthetized with isoflurane or awake and hooded. Radiographic images were obtained by use of a single digital system.b Focus distance was set at 100 cm, and exposure settings differed minimally among birds (65 to 70 kVp and 3.2 to 4.0 mA·s).
Radiographs were examined to ensure there was correct ventrodorsal alignment (ie, the keel and vertebrae were overlaid and there was symmetric extension of the limbs).1 To ensure measurements were comparable among birds, rotated views were excluded from the study.
Data analysis
There are conflicting reports regarding the reliability of sexing ospreys by examination of their plumage (females can have darker feathers).16,17 In addition, weight ranges for adult male and female ospreys overlap, which makes it difficult to determine sex on the basis of body weight.17 Therefore, differences between sexes were not considered in the study.
A significant difference in body weight between juvenile and adult female conspecifics has been reported.3 In addition, other investigators2,18 have reported an increase in relative cardiac size with an overall decrease in body weight. To reduce potential confounding attributable to age, birds in the study were limited to those with a body weight between 1.0 and 1.7 kg. Birds in the study were categorized as juvenile, adult, or unknown for statistical purposes.
Radiographic measurements for each bird were obtained by 3 of 4 observers (KMTW, ALD, SPK, and LABA), all of whom are veterinarians and were trained by and obtained measurements under the supervision of a board-certified avian specialist who had experience with radiography (GHB).
Single measurements for cardiac width, sternal width, and thoracic width were obtained from radiographs of each bird by the observers; observers were not aware of the operators who obtained radiographs to limit potential operator-observer bias. All measurements were obtained by use of digital software.c Cardiac width was measured at its widest point; sternal width and thoracic width were measured at that same level (Figure 1).
Ventrodorsal radiographic view of an osprey (Pandion haliaetus) illustrating measurements obtained for thoracic width (distance between points A), sternal width (distance between points B), and cardiac width (distance between points C). Measurements were obtained at the same level, which was based on the largest cardiac width.
Citation: American Journal of Veterinary Research 80, 9; 10.2460/ajvr.80.9.840
Statistical analysis
Two-way mixed-effects models were used to calculate intraclass correlation coefficients and interrater reliability, which assumed a fixed number of raters (n = 3) and no subject-by-rater interaction.19 Summary statistics (mean, SD, and 95% confidence interval) were calculated for each measure of interest. Bird-specific means were used to calculate ratios between cardiac width and sternal width and between cardiac width and thoracic width. The Anderson-Darling test was used to test for normal distribution of measured variables. Multivariable linear regression analysis was used to estimate the effect of sternal width and thoracic width (alone and in combination) on cardiac width while controlling for age (adult or juvenile). The model with the best fit was used to calculate theoretical cardiac width values, which were then plotted against the measured values. The SD for the model was derived by analyzing the difference between theoretical and measured cardiac width values and was used to create a theoretical reference range for cardiac width in healthy ospreys (within 2 SDs of the theoretical cardiac width).7 Values of P < 0.05 were considered significant. All analyses were conducted with statistical analysis software.d
Results
A total of 217 ospreys were admitted to the Clinic for the Rehabilitation of Wildlife between January 2015 and March 2018. Of these, 163 were excluded because of incomplete information in the medical record (n = 91), body weight < 1.0 kg or > 1.7 kg (22), misalignment on radiographs (16), physical examination findings (eg, heart murmur; 9), total solids concentration > 4 g/dL (9), and multiple criteria (16). Thus, the study comprised data for 54 birds (22 adults [6 females, 8 males, and 8 birds of unknown sex], 19 juveniles, and 13 birds of undetermined age). Body weight of the 54 ospreys ranged from 1.00 to 1.65 kg (mean ± SD, 1.34 ± 0.18 kg; median, 1.34 kg).
All measured variables were normally distributed. For the 54 ospreys, mean ± SD cardiac width was 37.19 ± 2.47 mm, mean sternal width was 40.73 ± 2.26 mm, and mean thoracic width was 54.65 ± 3.45 mm (Table 1). The mean cardiac width-to-sternal width ratio was 91%, and the mean cardiac width-to-thoracic width ratio was 68%. Interrater reliability was high for measurements of cardiac width (97%), sternal width (96%), and thoracic width (90%). Cardiac width was strongly correlated (r = 0.76) with sternal width and moderately correlated (r = 0.61) with thoracic width for all ospreys (Figure 2). These correlations differed slightly when juvenile and adult ospreys were compared.
Scatterplots of cardiac width versus sternal width (A through C) and cardiac width versus thoracic width (D through F) measured on radiographs obtained for 19 juvenile ospreys (A and D), 22 adult ospreys (B and E), and all 54 ospreys (combination of the juvenile and adult ospreys and 13 ospreys of undetermined age; C and F). The line in each panel represents the line of best fit for the available data. Notice that the scale on the x-axis in panels A through C differs from that in panels D through F.
Citation: American Journal of Veterinary Research 80, 9; 10.2460/ajvr.80.9.840
Values for variables measured on radiographs of 54 healthy ospreys (Pandion haliaetus) comprising 19 juveniles, 22 adults, and 13 birds of undetermined age.
| Variable | Minimum | Maximum | Mean | SD | 95% CI |
|---|---|---|---|---|---|
| Cardiac width (mm) | 31.92 | 44.08 | 37.19 | 2.47 | 36.52–37.87 |
| Sternal width (mm) | 35.18 | 45.93 | 40.73 | 2.26 | 40.11–41.34 |
| Thoracic width (mm) | 47.47 | 62.15 | 54.65 | 3.45 | 53.71–55.59 |
| Cardiac width-to-sternal width ratio (%) | 82.83 | 98.51 | 91.32 | 3.36 | 90.40–92.24 |
| Cardiac width-to-thoracic width ratio (%) | 59.89 | 75.85 | 68.13 | 3.47 | 67.18–69.07 |
CI = Confidence interval.
Individual linear regression models were used to separately assess the effect of sternal width and thoracic width on cardiac width, and multivariable linear regression analysis was used to model the effects of the combination of sternal width and thoracic width on cardiac width after controlling for age. Both sternal width and thoracic width were significant predictors of cardiac width, even after controlling for age (Table 2). Multivariable linear regression analyses resulted in the following model for predicting cardiac width:
Theoretical cardiac width = 1.1442 + (0.6423•sternal width) + (0.1709•thoracic width) + (0.8477•age)
Results of linear regression analysis for the prediction of theoretical cardiac width in healthy ospreys.
| Predictor variable | Estimate | 95% CI | P value* | R2 |
|---|---|---|---|---|
| Sternal width | 0.69 | |||
| Intercept | 0.02 | −6.71 to 6.82 | NS | |
| Sternal width | 0.91 | 0.75 to 1.08 | < 0.001 | |
| Thoracic width | 0.48 | |||
| Intercept | 10.10 | 2.37 to 17.9 | 0.01 | |
| Thoracic width | 0.50 | 0.35 to 0.64 | < 0.001 | |
| Combined | 0.72 | |||
| Intercept | 1.14 | −6.63 to 8.92 | NS | |
| Sternal width | 0.64 | 0.40 to 0.88 | < 0.001 | |
| Thoracic width | 0.17 | 0.02 to 0.32 | 0.03 | |
| Adult status | 0.85 | 0.05 to 1.64 | 0.04 |
Combined represents the combination of sternal width and thoracic width.
Values were considered significant at P < 0.05.
CI = Confidence interval. NS = Not significant.
where age is an indicator of adult or juvenile status (value = 1 for adult and 0 for juvenile).
Measured cardiac width was plotted against theoretical cardiac width (Figure 3). Mean ± SD difference between measured and theoretical values for cardiac width was 0.97 ± 0.75 mm. Measured cardiac width was considered within anticipated limits when the value was within 2 SDs of the calculated theoretical value. Thus, the theoretical reference range for an osprey was calculated as follows:
Theoretical cardiac width ± (2 × SD)
where SD = 0.75 mm. By use of this equation, measurements of cardiac width for 34 of 41 (83%) ospreys with known juvenile or adult status were within the theoretical reference range.
Scatterplot of measured cardiac width versus theoretical cardiac width in 41 ospreys of known age (19 juveniles and 22 adults). Theoretical cardiac width was calculated as 1.1442 + (0.6423•sternal width) + (0.1709•thoracic width) + (0.8477•age), where age is an indicator of adult or juvenile status (value = 1 for adult and 0 for juvenile). The line represents the line of best fit for the available data.
Citation: American Journal of Veterinary Research 80, 9; 10.2460/ajvr.80.9.840
Discussion
Results of the study reported here may provide a standard by which ospreys can be assessed radiographically for cardiomegaly. Healthy juvenile and adult ospreys had a cardiac width that was 90% to 92% of the sternal width and 67% to 69% of the thoracic width as measured on ventrodorsal radiographic images. In addition, cardiac width may be predicted by use of an equation based on measured sternal width and thoracic width (ie, theoretical cardiac width = 1.1442 + [0.6423•sternal width] + [0.1709•thoracic width] + [0.8477•age]), which would allow clinicians to compare values for measured and predicted cardiac width to aid in identifying cardiomegaly.
For example, measurements for an adult osprey (thoracic width, 61 mm; sternal width, 42 mm; and cardiac width, 40 mm) could be used to calculate the theoretical cardiac width as follows: theoretical cardiac width = 1.1442 + (0.6423•42) + (0.1709•61) + (0.8477•1) = 39.39 mm. The reference range for the theoretical cardiac width for a bird with these measurements would be 39.39 ± (2 × 0.75) = 39.39 ± 1.50 mm (ie, 37.89 to 40.89 mm), which would indicate that the cardiac width for that osprey was within reference limits. To our knowledge, the study reported here was the first in which radiographic standards of cardiac width for healthy juvenile and adult ospreys have been determined.
Investigators have evaluated the correlation between cardiac width and thoracic width in multiple species of birds and found that thoracic width is a significant predictor of cardiac width.2–8,a However, thoracic width can be influenced by the phase of the respiratory cycle during which a radiograph is obtained.6,a Investigators for most of the studies did not consider the phase of respiration when radiographs were obtained, which may limit validity of the estimates for cardiac width-to-thoracic width ratio.
Sternal width was also found to be a significant predictor of cardiac width.6,7,a In the present study, both sternal width and thoracic width were significant predictors of cardiac width (sternal width was more strongly correlated with cardiac width than was thoracic width). In addition, when controlling for age (juvenile or adult), the resulting multivariate equation better explained the associations among sternal width, thoracic width, and cardiac width.
Cardiac width has also been compared with coracoid width.2,5,7,8,a A strong correlation2,8 or a moderate correlation5 was detected between the 2 measurements. In the other studies,7,a investigators found no significant correlation between cardiac width and coracoid width, possibly because of the small size of the coracoid, which can contribute to measurement errors.2 Coracoid width was not measured in the present study because of the possibility of measurement errors as well as variability in the correlation between coracoid width and cardiac width.
Additional radiographic measures used to predict cardiac width include clavicular width,4,8 rib space length,4,5,8 sternal length,2,5 cardiac length,2 and sacrum width.4 A correlation was detected between cardiac length and sternal length in one study5 and between cardiac width and sternal length in another study.2 In those studies, lateral radiographic images were obtained as part of a routine diagnostic evaluation; however, sternal length and cardiac length are not always evident radiographically owing to superimposition of the sternum with the proventriculus and cardiac apex with the liver2,4–6; thus, those measurements were not included in the present study. Cardiac length and sternal length have been measured by manipulating images5 or by direct measurements of birds obtained during physical examination.2 Because of the retrospective nature of the study, and in an attempt to eliminate possible subjective measurements that often can be difficult to obtain, measurements on lateral radiographic views were not obtained in the present study.
Although diagnostic tests such as echocardiography, CT, and MRI may be ideal for use in diagnosing cardiac disease, these modalities are not widely used in wildlife medicine because of the advanced skills and specialized equipment required and the cost for consultation with specialists. Therefore, these tests were not included in the present study. However, other noninvasive diagnostic tests may prove to be a useful alternative in cardiac assessment of ospreys and other avian species. There is an association between cardiac troponin I concentrations and cardiac damage in double-crested cormorants.20 Reference limits for diagnostic laboratory tests (eg, cardiac bio-markers) could complement the results reported here and aid clinicians in assessment of cardiac health of ospreys.
One of the limitations of the study reported here was the possibility of subclinical cardiac disease that did not result in observable illness or clinical signs, which would have potentially biased the reported estimates (ie, larger values). Additionally, few healthy ospreys are admitted to wildlife hospitals unless they are found as juveniles or misplaced during a storm, so there may have been bias in the present study.
Another limitation was the potential effect of sex and body weight on cardiac measures. Ospreys may differ in body weight on the basis of geographic location attributable to migration. Ospreys are present in Florida throughout the year, but they migrate to other regions of the United States. Juvenile ospreys are not dimorphic, and there may be some overlap in adult plumage, which may cause results of sexing to be inaccurate. In the study reported here, sex was reported for 41 ospreys; because of this small number, differences between sexes could not be evaluated as a confounding factor. Sex may be determined through laboratory testing, but the cost may be prohibitive for wildlife clinics; therefore, sex was not determined in this manner. Furthermore, sex may affect the physical fitness status and, therefore, cardiac width. Additional studies are needed to determine the manner in which differences in sex, age, and body size impact cardiac width.
Finally, radiographic measurements were obtained by observers under the supervision of a board-certified avian specialist by use of 1 digital program to limit bias. However, this would not fully prevent interobserver differences and must be considered a limitation to the study; veterinarians making measurements in field settings will have interobserver differences on the basis of the use of a variety of digital programs as well as human error.
Ospreys are often used as sentinel species to monitor the health of ecosystems. Despite their importance in this role, there is a paucity of information on standard measures for ospreys on which to base potential deviations. The study reported here provided a standard by which juvenile and adult ospreys can be evaluated radiographically for cardiomegaly, which increases their usefulness as sentinels in native habitats. Further studies are warranted to develop more robust standards for evaluating osprey health and, consequently, the health of local environments.
Acknowledgments
The authors did not receive any extrainstitutional funding or support for this study and have no conflicts of interest to report.
Footnotes
Lecona IG, Barncord K, Stauthammer C. Cardiac evaluation in the bald eagle (Haliaeetus leucocephalus) (abstr), in Proceedings. ExoticsCon 2015;143–144.
Vet-Ray Digital Vet DX system, Sedecal, Buffalo Grove, Ill.
QXVue software, Sedecal, Buffalo Grove, Ill.
SAS, version 9.4, SAS Institute, Cary, NC.
References
1. McMillan M. Imaging techniques. In: Ritchie BW, Harrison GJ, Harrison LR, eds. Avian medicine: principles and applications. Lake Worth, Fla: Wingers Publishing, 1994;246–326.
2. Straub J, Pees M, Krautwald-Junghanns ME. Measurement of the cardiac silhouette in psittacines. J Am Vet Med Assoc 2002;221:76–79.
3. Rettmer H, Watson R, Hatt JM, et al. Radiographic measurement of internal organs in Spix's macaws (Cyanopsitta spixii). J Avian Med Surg 2011;25:254–258.
4. Velayati M, Mirshahi A, Razmyar J, et al. Radiographic reference limits for cardiac width of budgerigars (Melopsittacus undulatus). J Zoo Wildl Med 2015;46:34–38.
5. Barbon AR, Smith S, Forbes N. Radiographic evaluation of cardiac size in four falconiform species. J Avian Med Surg 2010;24:222–226.
6. Lumeij JT, Shaik MAS, Ali M. Radiographic reference limits for cardiac width in peregrine falcons (Falco peregrinus). J Am Vet Med Assoc 2011;238:1459–1463.
7. Hanley CS, Murray HG, Torrey S, et al. Establishing cardiac measurement standards in three avian species. J Avian Med Surg 1997;11:15–19.
8. Mirshahi A, Shariatzaden M, Razmyar J, et al. Evaluation of cardiac size in the common kestrel (Falco tinnunculus) based on radiographic measurements. J Avian Med Surg 2016;30:345–349.
9. Klasing KC, Dierenfeld ES, Koutsos EA. Avian iron storage disease: variations on a common theme? J Zoo Wildl Med 2012;43:S27–S34.
10. Niemeyer C, Favero CM, Kolesnikovas CKM, et al. Two different avipoxviruses associated with pox disease in Magellanic penguins (Spheniscus magellanicus) along the Brazilian coast. Avian Pathol 2013;42:546–551.
11. Vanstreels RET, Kolesnikovas CKM, Sandri S, et al. Outbreak of avian malaria associated to multiple species of Plasmodium in Magellanic penguins undergoing rehabilitation in Southern Brazil. PLoS One 2014;9:e94994.
12. Sharma D, Asrani RK, Ledoux DR, et al. Toxic interaction between fumonisin B1 and moniliformin for cardiac lesions in Japanese quail. Avian Dis 2012;56:545–554.
13. United States Geological Survey. Integrated taxonomic information system on-line database. Pandion haliaetus. Available at: www.itis.gov. Accessed Mar 29, 2019.
14. United States Geological Survey. Longevity records of North American birds. Available at: www.pwrc.usgs.gov/bbl/longevity/longevity_main.cfm. Accessed Mar 29, 2019.
15. Grove RA, Henny CJ, Kaiser JL. Osprey: worldwide sentinel species for assessing and monitoring environmental contamination in rivers, lakes, reservoirs, and estuaries. J Toxicol Environ Health Part B Crit Rev 2009;12:25–44.
16. Cornell Laboratory of Ornithology. Osprey identification. Available at: www.allaboutbirds.org/guide/Osprey/id. Accessed Dec 17, 2018.
17. Virginia Institute of Marine Science. How to identify an osprey. Available at: www.vims.edu/bayinfo/ospreycam/about/identification/index.php. Accessed Dec 17, 2018.
18. Lumeij JT, Ritchie BW. Cardiology. In: Ritchie BW, Harrison GJ, Harrison LR, eds. Avian medicine: principles and application. Lake Worth, Fla: Wingers Publishing, 1994;695–722.
19. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86:420–428.
20. Daigneault M, Harr KE, Dean KM, et al. Assay validation of the cardiac isoform of troponin I in double crested cormorant (Phalacrocorax auritus) plasma for diagnosis of cardiac damage. Ecotoxicol Environ Saf 2017;141:52–56.
