Anticoagulant Rodenticide Exposure in South Carolina Raptors

Patterns and Clinical Outcomes, 2024-2025

Authors
Affiliations

Valerie Sprinkel

Medical Clinic Technician, Avian Conservation Center

Julian L. Costa

Statistician

Published

October 6, 2025

Data Coverage & Completeness

This dataset includes 48 raptors tested between 2024–2025.

  • Sampling: Most (81%) of the submissions were from 2025 (39 of 48).

  • Species: Great Horned Owls make up the majority (60%), followed by Red-tailed Hawks (23%). Other species were present in smaller numbers.

  • Completeness: Key fields such as coordinates (98%), sex (83%), weight (83%), and clinical/necropsy notes (88–96%) are well recorded overall.

  • Age/Sex: About half of birds had age noted, while sex was recorded in most cases. Only a few entries (4%) were flagged as “sex assumed.”

  • Rodenticide detection: Nearly 90% of birds tested positive for at least one compound.

Table 1. Number of Raptors Sampled by Year
year Number of Samples
2024 9
2025 39
Table 2. Distribution of Bird Species
species_name Count % of Sample
Great Horned Owl 29 60.4
Red-tailed Hawk 11 22.9
Barred Owl 3 6.2
Red-shouldered Hawk 2 4.2
Barn Owl 1 2.1
Osprey 1 2.1
Turkey Vulture 1 2.1

Table 3. Data Completeness by Key Field
Field Percent
With Coordinates 97.9
With Sex 83.3
Sex Assumed 4.2
With Weight 83.3
With Exam Notes 95.8
With Necropsy 87.5
Table 4. Age/Sex Field Coverage
Known Age (%) Known Sex (%) Sex Assumed (%)
54.2 83.3 4.2
Table 5. Age × Sex × Assumed Breakdown
Age Class Sex Sex Assumed Count
Adult Male FALSE 11
NA Male FALSE 10
Adult Female FALSE 9
NA NA FALSE 7
NA Female FALSE 5
Juvenile Female FALSE 2
Adult Female TRUE 1
Adult Male TRUE 1
Adult NA FALSE 1
Hatchling Male FALSE 1
Table 6. Any Rodenticide Detected (Bird-level)
Status n Percent
No compounds detected 5 10.4
Detected ≥1 compound 43 89.6

Geography & Potential Hotspots

Admissions were mapped across South Carolina counties. Most admitted raptors came from coastal and more populated areas, while very few were admitted from upstate.

  • Multi-compound detections: Found in both coastal and inland counties. Birds admitted from coastal areas were significantly more likely to show multiple compounds (Fisher’s Exact Test, p = 0.034, odds ratio ≈ 6.8). However, the confidence interval is wide, so this should be interpreted cautiously.

  • County-level view: Coastal counties had more admitted birds overall.

  • Note: Geographic results reflect where birds were admitted from, which may not match the actual exposure location. Patterns could also be influenced by factors such as human reporting or transport, not just where exposures occur.

Table 7. Multi-Compound Detections by County
County Birds Multi-Compound Rate (%)
Allendale 1 1 100.0
Berkeley 3 3 100.0
Colleton 1 1 100.0
Florence 1 1 100.0
Charleston 17 14 82.4
Beaufort 3 2 66.7
Horry 3 2 66.7
Georgetown 2 1 50.0
Clarendon 1 0 0.0
Dorchester 2 0 0.0
Orangeburg 1 0 0.0
Williamsburg 1 0 0.0
NA 2 0 0.0
Table 8. Multi-Compound Rate by Region (Coastal vs Non-coastal)
Region N Rate (%)
Coastal 31 74.2
Non-coastal 7 28.6 

    Fisher's Exact Test for Count Data

data:  coastal_mat
p-value = 0.03413
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
  0.8940933 84.1701147
sample estimates:
odds ratio 
  6.757918

Dose vs Clinical/Necropsy Findings

Clinical Severity vs Dose

We compared rodenticide burden between birds with severe clinical findings and those with milder findings.

  • Mild/Other cases (n = 22): mean 1.47 µg/g, median 0.17, max 7.36 µg/g.

  • Severe cases (n = 24): mean 3.45 µg/g, median 0.60, max 23.5 µg/g.

  • Statistical test: Wilcoxon rank-sum test → not significant (W = 243, p = 0.65).

  • Interpretation: Severe cases tended to have higher burdens on average, but wide overlap suggests that dose alone does not predict severity.

Table 9. Total Dose (µg/g) by Clinical Severity
exam_severity_level n mean_dose median_dose sd_dose min_dose max_dose
Mild/Other 22 1.47 0.17 2.10 0 7.36
Severe 24 3.45 0.60 6.04 0 23.52 


    Wilcoxon rank sum test with continuity correction

data:  total_dose_ug_g by exam_severity_level
W = 243, p-value = 0.652
alternative hypothesis: true location shift is not equal to 0

Hemorrhage Necropsy Findings vs Dose

We compared rodenticide burden between birds with and without hemorrhage-related necropsy findings.

  • Hemorrhage absent (n = 15): mean 3.10 µg/g, median 0.14, max 23.5 µg/g.

  • Hemorrhage present (n = 27): mean 2.33 µg/g, median 0.89, max 18.6 µg/g.

  • Statistical test: Wilcoxon rank-sum test → not significant (W = 174.5, p = 0.47).

  • Interpretation: Birds with hemorrhage tended to have higher median rodenticide burdens, though those without hemorrhage included some of the highest observed outliers. Overall, the difference between groups was not statistically significant.

Table 10. Total Dose (µg/g) by Hemorrhage Necropsy Finding
hemorrhage_flag n mean_dose median_dose sd_dose min_dose max_dose
Hemorrhage: No 15 3.10 0.14 6.23 0 23.52
Hemorrhage: Yes 27 2.33 0.89 3.87 0 18.60 


    Wilcoxon rank sum test with continuity correction

data:  total_dose_ug_g by hemorrhage_flag
W = 174.5, p-value = 0.4703
alternative hypothesis: true location shift is not equal to 0

High-Dose Severe Cases

We reviewed the 10 highest-dose birds that were classified as severe cases. These admissions illustrate the clinical and necropsy signs most often associated with heavy rodenticide exposure.

  • Doses ranged from 1.69 µg/g to 23.52 µg/g.

  • Several birds showed classic coagulopathy signs (e.g., hemorrhage, shock, collapse).

  • Multiple birds (n = 3) died within 24 hours of admission.

  • Necropsy findings frequently documented hemorrhage, bruising, or organ damage.

Table 11. Highest-Dose Birds (Severe Cases)
Year Species Total Dose (µg/g) # Compounds Weight (g) Exam Notes Necropsy Notes
2025 Red-tailed Hawk 23.5189 4 1200 Coagulopathy; Weakness/collapsed chasing prey Crop hemorrhage
2024 Great Horned Owl 18.6000 3 1200 Shock; Died within 24 hours; Laying prone Blood unknown source
2025 Great Horned Owl 9.3500 4 1492 Shock; Soft tissue wounds Bruising other
2024 Great Horned Owl 6.8000 3 1046 Bird on ground, unable to fly; Dehydration; Hypovolemic; Died within 24 hours; Shock NA
2025 Red-tailed Hawk 5.6600 2 1222 Emaciated; No deep pain response Bruising chest; Cholecystitis; Hemopneumopericardium / effusion / retrosternal hemothorax
2024 Red-tailed Hawk 5.2000 2 1314 Bird on ground, unable to fly; Lethargic; Suspect hit by car; Died in 4 days Bruising abdomen; Crop hemorrhage
2025 Great Horned Owl 3.5080 3 1041 Died within 24 hours; Shock Pericardial hemorrhage
2025 Red-tailed Hawk 3.3500 3 1040 Suspect hit by car; Hypovolemic; Eye hemorrhage Bruising abdomen; Hemopneumopericardium / effusion / retrosternal hemothorax
2025 Great Horned Owl 2.4293 2 1518 Bird on ground, unable to fly; Pupillary hippus; Vent tonal laxity Bruising chest; Hepatitis
2025 Great Horned Owl 1.6860 3 1198 Coagulopathy; Hypovolemic; Soft tissue wounds Bruising abdomen; Hemoperitoneum; Hemopneumopericardium / effusion / retrosternal hemothorax; Pulmonary hemorrhage; Biliverdin jaundice

Weight vs Dose vs Symptoms

We looked at rodenticide dose relative to bird weight to see if smaller or larger raptors showed different vulnerability. By calculating a weight-adjusted dose (µg of toxin per gram of body weight), we highlight which birds carried disproportionately high toxic loads.

Table 12 lists the 10 birds with the highest weight-adjusted burdens. While most of these were large-bodied raptors, we note that one smaller Great Horned Owl (690 g) also appeared in the list.

Across these top cases, birds with the highest adjusted doses frequently showed severe clinical signs, including coagulopathy, collapse, hemorrhage, and in several instances, death within 24 hours.

Table 12. Highest Weight-Adjusted Dose Birds
Year Species Total Dose (µg/g) Weight (g) Burden per Gram Exam Notes Necropsy Notes
2025 Red-tailed Hawk 23.5189 1200 0.0195991 Coagulopathy; Weakness/collapsed chasing prey Crop hemorrhage
2024 Great Horned Owl 18.6000 1200 0.0155000 Shock; Died within 24 hours; Laying prone Blood unknown source
2024 Great Horned Owl 6.8000 1046 0.0065010 Bird on ground, unable to fly; Dehydration; Hypovolemic; Died within 24 hours; Shock NA
2025 Great Horned Owl 9.3500 1492 0.0062668 Shock; Soft tissue wounds Bruising other
2025 Great Horned Owl 5.6000 1100 0.0050909 Attacked by crows; Died within 24 hours Bruising chest; Cholecystitis; Hemorrhagic gizzard; Hemopneumopericardium / effusion / retrosternal hemothorax
2025 Red-tailed Hawk 5.6600 1222 0.0046318 Emaciated; No deep pain response Bruising chest; Cholecystitis; Hemopneumopericardium / effusion / retrosternal hemothorax
2025 Great Horned Owl 3.0000 690 0.0043478 Emaciated; Hypothermic; Bird on ground, unable to fly; Died within 24 hours Bruising abdomen; Bruising other; Cholecystitis; Biliverdin jaundice
2024 Red-tailed Hawk 5.2000 1314 0.0039574 Bird on ground, unable to fly; Lethargic; Suspect hit by car; Died in 4 days Bruising abdomen; Crop hemorrhage
2025 Great Horned Owl 3.5080 1041 0.0033698 Died within 24 hours; Shock Pericardial hemorrhage
2025 Red-tailed Hawk 3.3500 1040 0.0032212 Suspect hit by car; Hypovolemic; Eye hemorrhage Bruising abdomen; Hemopneumopericardium / effusion / retrosternal hemothorax

Species vs Clinical/Necropsy Findings

Species-level patterns were summarized for admissions with adequate sample size (n ≥ 5). Among species with at least five birds, we report the percentage with severe clinical findings and the percentage with hemorrhage noted at necropsy.

Table 13. Species with ≥ 5 Admissions: % Severe and % Hemorrhage
Species N Severe (n) Severe (%) Hemorrhage (n) Hemorrhage (%)
Great Horned Owl 29 14 48.3 16 55.2
Red-tailed Hawk 11 7 63.6 7 63.6

 

A Fisher’s Exact Test evaluated whether rates of severe findings differed by species.

Clinical Severity vs Species

Fisher’s Exact Test showed no significant difference in the proportion of severe cases across species (p = 0.72), suggesting that no species clearly stood out as more likely to present with severe clinical signs.

  • Odds ratio ≈ 1.6 (95% CI: 0.32–9.33).

    Fisher's Exact Test for Count Data

data:  sev_tab
p-value = 0.721
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.3154699 9.3306068
sample estimates:
odds ratio 
    1.6045

Hemorrhage Findings vs Species

Fisher’s Exact Test showed no significant difference in hemorrhage-related necropsy findings across species (p = 1.00), indicating that hemorrhage findings were distributed fairly evenly across species.

  • Odds ratio ≈ 1.16 (95% CI: 0.19–8.84).

    Fisher's Exact Test for Count Data

data:  hemo_tab
p-value = 1
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.191881 8.844297
sample estimates:
odds ratio 
  1.161448

Sex vs Clinical/Necropsy Findings

We also tested whether clinical severity and hemorrhage at necropsy differ by sex.

  • Severe cases: More common in females (71%) than males (39%), though results suggested that there was no statistical difference (Fisher’s Exact Test, p = 0.721).

  • Hemorrhage findings: Rates were similar between sexes (59% females vs 61% males), with no statistical difference (Fisher’s Exact Test, p = 1.00).

These results suggest that while there may be a tendency for females to present more often with severe findings, hemorrhage was equally observed in both sexes.

Table 14. Sex vs Clinical/Necropsy Findings (Known Sex Only)
Sex N Severe (n) Severe (%) Hemorrhage (n) Hemorrhage (%)
Male 23 9 39.1 14 60.9
Female 17 12 70.6 10 58.8 

    Fisher's Exact Test for Count Data

data:  sev_tab
p-value = 0.721
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.3154699 9.3306068
sample estimates:
odds ratio 
    1.6045 

    Fisher's Exact Test for Count Data

data:  hemo_tab
p-value = 1
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.1899573 4.9948515
sample estimates:
odds ratio 
         1

Compound-Specific Patterns

We examined which anticoagulant rodenticides (ARs) were most commonly detected across all raptor admissions.

  • Dominant compound: Brodifacoum was the most common AR, detected in 81% of birds.

  • Other frequent SGARs: Difethialone (54%) and bromodiolone (50%) were also widely present.

  • Less common SGARs: Difenacoum was detected in 8% of birds.

  • FGARs: Diphacinone was the only FGAR with notable detections (17%), while chlorophacinone appeared in just one bird (2%). Warfarin, coumachlor, coumatetralyl, and dicoumerol were not detected.

  • SGAR vs FGAR: Nearly 88% of birds had at least one SGAR present, compared with only 19% showing any FGAR detection.

Table 15. SGAR vs FGAR Detection Prevalence
SGAR Positive (n) FGAR Positive (n) SGAR (%) FGAR (%)
42 9 87.5 18.8
Table 16. Prevalence of Each Anticoagulant Rodenticide (AR) Across Birds
Compound Positive Birds Prevalence (%)
brodifacoum 39 81.2
difethialone 26 54.2
bromodiolone 24 50.0
diphacinone 8 16.7
difenacoum 4 8.3
chlorophacinone 1 2.1
coumachlor 0 0.0
coumatetralyl 0 0.0
dicoumerol 0 0.0
warfarin 0 0.0

Table 16B. Concentration Summary by Compound (µg/g)
Compound Positive Birds Mean Median Max
brodifacoum 39 1.8 0.2 23.4
diphacinone 8 1.57 0.01 8.4
bromodiolone 24 0.95 0.46 6.7
difethialone 26 0.41 0.09 2.6
difenacoum 4 0.25 0.01 1
chlorophacinone 1 7e-05 7e-05 7e-05
coumachlor 0 NaN NA 0
coumatetralyl 0 NaN NA 0
dicoumerol 0 NaN NA 0
warfarin 0 NaN NA 0

Toxin Type vs Findings (SGAR/FGAR)

We tested whether the presence of any SGAR or FGAR was associated with clinical severity or hemorrhage.

SGARs

Birds with ≥1 SGAR were somewhat less likely to be classified as severe (48% vs 67%), but this was not statistically significant (p = 0.35). Hemorrhage was more common when SGARs were present (62% vs 17%), but this also was not statistically significant (p = 0.12).

Table 20. Any SGAR Present vs Clinical/Necropsy Findings
SGAR Present N Severe (n) Severe (%) Hemorrhage (n) Hemorrhage (%)
No SGAR 6 4 66.7 1 16.7
Yes (≥1 SGAR) 42 20 47.6 26 61.9 

    Fisher's Exact Test for Count Data

data:  sev_sgar_tab
p-value = 0.3488
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.004624645 2.756775534
sample estimates:
odds ratio 
 0.2448678 

    Fisher's Exact Test for Count Data

data:  hemo_sgar_tab
p-value = 0.122
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
   0.4452819 353.1803346
sample estimates:
odds ratio 
  6.195002

FGARs

Results were based on small numbers (n = 9 with any FGAR). Severe cases (44%) and hemorrhage (67%) were observed at similar rates to birds without FGARs, and no significant differences were found (p ≈ 0.7 for both tests).

Table 21. Any FGAR Present vs Clinical/Necropsy Findings (Sparse)
FGAR Present N Severe (n) Severe (%) Hemorrhage (n) Hemorrhage (%)
No FGAR 39 20 51.3 21 53.8
Yes (≥1 FGAR) 9 4 44.4 6 66.7 

    Fisher's Exact Test for Count Data

data:  sev_fgar_tab
p-value = 0.7178
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
 0.1160849 3.7642992
sample estimates:
odds ratio 
 0.6857591 

    Fisher's Exact Test for Count Data

data:  hemo_fgar_tab
p-value = 0.6888
alternative hypothesis: true odds ratio is not equal to 1
95 percent confidence interval:
  0.2712776 21.2100202
sample estimates:
odds ratio 
  1.831581

 

Overall, SGAR exposures were more frequent and showed a tendency toward higher hemorrhage rates, but neither SGAR nor FGAR presence was significantly linked to severity or hemorrhage in this sample.

Summary & Conclusions

This analysis examined anticoagulant rodenticide (AR) exposure patterns in 48 South Carolina raptors based on toxicology testing coordinated by the Avian Conservation Center (Awendaw, SC) during 2024-2025.

Key Findings

  • High prevalence of exposure: Nearly 90% of raptors tested positive for at least one anticoagulant rodenticide (AR). Brodifacoum, difethialone, and bromodiolone were the most common compounds detected.

  • Widespread co-exposures: Over 60% of birds carried multiple AR compounds, with significantly higher multi-compound rates in coastal counties.

  • Clinical patterns: Birds with severe clinical or necropsy findings generally carried higher rodenticide burdens, but differences were not statistically significant. Hemorrhage findings were common, though not uniquely tied to higher doses.

  • Demographics: Severe findings were more frequent in females (71%) compared to males (39%), though species-level comparisons did not reveal meaningful differences.

  • Geography and timing: Coastal regions accounted for most (74%) of the admissions and showed greater multi-compound detections. Patterns were consistent across 2024 and 2025, with no major shifts in compound profiles.

Recommendations for Future Work

  • Expand sample size and diversity: Increasing the number of admissions tested, and including a wider variety of species, would strengthen analyses and improve statistical power.

  • Integrate ecological context: Combine toxicology results with prey availability, land use, and rodenticide application data to better understand exposure pathways.

  • Collaborative expansion: Partner with wildlife centers across the Southeast to broaden data collection and generate more robust regional assessments.

While these findings mark important progress in understanding rodenticide exposure, they also underscore the need for continued awareness and expanded research to protect raptor populations into the future.