This case involved a 2-year-old 46 kg healthy female Irish Wolfhound who received routine ovariohysterectomy and prophylactic gastropexy surgical procedures. Medical history was unremarkable other than experiencing a small ear laceration from a dog bite at 7 months of age. This was initially treated conservatively by bandaging but required surgical repair by suturing 1 week after the initial trauma because of persistent bleeding from the wound. No abnormalities were detected on a routine physical examination prior to the current surgery. Presurgical hematology and chemistry panel values were within the reference ranges for the laboratory other than a slightly decreased white blood cell count (4.6 × 10³ / μL, reference range 4.9–17.7 × 10³ / μL) and a slightly increased hematocrit (57.5%, reference range 38.3–56.5%).

The dog was premedicated with hydromorphone (0.15 mg kg⁻¹, SC), maropitant (1 mg kg⁻¹, SC), and midazolam (0.2 mg kg⁻¹, IV), and anesthesia induced with propofol (4 mg kg⁻¹, IV). The trachea was intubated, and anesthesia maintained with isoflurane in oxygen. Carprofen (2.2 mg kg⁻¹, SC) was administered 30 min after induction. The dog was placed in dorsal recumbency and the abdominal skin clipped and then disinfected using chlorhexidine scrub and solution.

The gastropexy was performed through a 20 cm incision caudal to xyphoid to mid abdomen. A 1.5 cm incision was made through the serosa and initial muscle layer of the 10th rib midway between dorsum and ventrum. A matching 1.5 cm incision was made in the serosa of the cranial surface of the stomach near the greater curvature, avoiding superficial blood vessels. The cranial margin of the stomach incision was sutured to the cranial margin of the body wall incision, and the caudal margin of the stomach incision was sutured to the caudal margin of the body wall incision, each with 3–0 poliglecaprone 25 absorbable monofilament sutures, in a simple continuous pattern.

The ovariohysterectomy was performed next. The uterus was isolated and the left ovary was exteriorized for visualization after stretching of the suspensory ligament. Two hemostats were placed across the ovarian pedicle. The clamp farthest from the ovary was then replaced with a size 2–0 poliglecaprone monofilament ligature tied using a Miller’s knot. The pedicle was then transected below the remaining clamp and checked for evidence of bleeding before release. The same technique was used for the right ovary. The cervix was then exposed, and a similar 2 clamp technique was used on the uterine body proximal to the cervix to place a size 2–0 poliglecaprone monofilament ligature tied with a Miller’s knot. The uterine body was transected below the clamp and the stump inspected for bleeding when returned to a normal anatomic position. After inspecting the abdomen for signs of hemorrhage, the linea alba was closed using size 0 polydioxanone sutures in a cruciate pattern. The subcutaneous tissues and skin were closed using 2–0 poliglecaprone monofilament sutures.

Recovery from anesthesia was unremarkable. The dog was discharged to the owner that evening with instructions for post-surgical care, including administration of carprofen tablets (2.2 mg kg⁻¹ orally) and gabapentin capsules (4.4 mg kg⁻¹ orally) every 12 h. For the first 2 days after surgery the owner reported that the dog was relatively quiet and appeared somewhat depressed, which was attributed to the sedative effects of gabapentin. Although the dog was much brighter on Day 3, the owner noted extensive purple discoloration of the skin (purpura) on the ventral abdomen (Figure 1A) that spread to the torso on Day 4 (Figure 1B). On the evening of Day 4, the dog was observed by the owner to collapse, struggle momentarily as if experiencing a seizure, and died. The owner brought the dog to an emergency clinic and requested a necropsy.

A gross necropsy examination was performed by the Nevada Department of Agriculture Animal Disease Laboratory. Major findings included significant skin discoloration (dark red/purple) over the entire ventral abdomen and thorax. Extensive subcutaneous hemorrhages up to 3 cm deep on the cut surface were noted in the same areas. Visible hemorrhages were also observed on the pleural surfaces of the thoracic wall, the serosal surface of the lower trachea, epicardial surface of the heart, the serosal surface of the urinary bladder, and scattered throughout the mesentery. The abdominal cavity contained 100 mL of free blood and a large (41 gram) blood clot. The stomach was sutured to the left abdominal wall, which was consistent with a prophylactic gastropexy. All ligatures were intact and secure on both ovarian pedicles and the uterine stump. Samples were collected from liver, heart and lungs for aerobic bacterial culture. No pathogens were subsequently identified. The pathologist suspected DEPOH and a sample of liver was submitted for SERPINF2 c.605 C > T genotyping.

A custom allele discrimination assay (Applied Biosystems TaqMan SNP Genotyping Assay, Thermo Fisher Scientific, Waltham MA) was used to genotype DNA samples for the SERPINF2 c.605 C > T gene variant as previously described in detail (8). Briefly, primers and probes were 5’-ACG CTG CGG AGG TTA GAG-3′ (forward primer), 5’-CCC AGG TCC TGG CAA AGG-3′ (reverse primer), 5’-CCA GAG TCT GCA TGC AG-3′ (C-allele probe labeled with VIC dye), and 5’-CCA GAG TCT ACA TGC AG-3′ (T-allele probe labeled with FAM dye). Assays were performed according to the manufacturer’s directions using a real-time PCR instrument (CFX96 Touch, Bio-Rad, Hercules, CA). The PCR method involved an initial denaturation step at 95 °C for 10 min, and 55 cycles of 95°C for 15 s and 60 °C for 90 s. Genotype calls were made using the CFX90 allele discrimination function. Positive and negative controls were included with each run. Positive control samples were randomly selected dog DNA samples with known SERPINF2 c.605 C > T genotype (C/C, C/T and T/T). The negative control sample was DNA-free pure water.

Stored DNA samples (N = 4,044) from client-owned dogs were retrieved from the Washington State University Veterinary Teaching Hospital Patient DNA Bank and the Comparative Pharmacogenomics Laboratory Sighthound DNA Bank. DNA had been extracted from buccal swabs or whole blood samples provided by the dog’s owner. Hospital patient samples were from dogs living in the Pacific Northwest of the United States, while the Sighthound DNA bank samples were obtained primarily by mail from dogs living throughout the United States. The dog’s breed was based on breed registration information (if available) and by owner designation. The 4,044 DNA samples represented 63 different dog breeds (16 Sighthound breeds and 47 other breeds) with a minimum of 20 DNA samples per breed and 186 DNA samples from mixed-breed dogs. The designation of a breed as belonging to the Sighthound group was primarily based on the breed inclusion criteria used for Lure Coursing established by the American Kennel Club¹. The collection, storage, and use of the DNA samples used in this study were approved by the Institutional Animal Care and Use Committee at Washington State University (protocols #04194 and #04539).

SERPINF2 c.605 C > T genotype data were extracted from publicly available whole genome sequences for 315 unrelated dogs from 50 different dog breeds (13). SERPINF2 c.605 C > T genotypes were also obtained from the same database for 27 unrelated gray wolves (Canis lupus) sampled from populations in North America, Europe and Asia.

The online tool PolyPhen-2 (Polymorphism Phenotyping v2)², was used to evaluate the possible functional impact of the SERPINF2 c.605 C > T variant, which results in an alanine to valine substitution at position 202 in the canine alpha-2 antiplasmin polypeptide. For this analysis, the reference canine alpha-2 antiplasmin amino acid sequence (XP_005624992.1) was retrieved from the National Center for Biotechnology Information protein sequence repository³.

SERPINF2 c.605 C > T genotyping indicated that the DEPOH case had the SERPINF2 c.605 T/T genotype (homozygous variant).

Canine population genotype frequencies and breed-specific prevalence of the SERPINF2 c.605 C > T variant were determined by genotyping DNA samples collected from 63 different breeds and supplemented using publicly available dog whole genome sequence data (13). At least 20 dogs per breed were sampled. Breeds evaluated included 16 sighthound breeds, 47 other (non-sighthound) dog breeds, and 186 mixed-breed dogs (4,359 dogs genotyped in total). SERPINF2 c.605 C > T genotype data for 27 gray wolves sampled from geographically diverse populations were also obtained from the same public database (13) to determine the ancestral origin of this variant.

Supplementary Table 1 provides the numbers of individual animals with each SERPINF2 c.605 genotype (C/C, C/T or T/T) as well as the variant allele frequencies (number of variant T alleles expressed as a percentage of the total number of alleles) for each dog breed, mixed-breed dogs and wolves. Figure 2 shows the percentage of canines within each group with the SERPINF2 c.605 homozygous variant (T/T) genotype and the heterozygous variant (C/T) genotype (8).

Greyhounds had the highest variant (T) allele frequency (52%) of all dog breeds evaluated with a high-risk T/T genotype prevalence of 29%. Irish Wolfhound (51%), Basenji (49%) and Italian Greyhound (42%) breeds also showed a high allele prevalence (frequencies over 35%). T/T genotype prevalence in these breeds ranged from 21 to 24%. Galgo Español (34%), Saluki (30%) Whippet (25%), Scottish Deerhound (23%), and Peruvian Inca Orchid (14%), showed a moderate allele prevalence (frequencies between 10 and 35%). T/T genotype prevalence in these breeds ranged from 5 to 13%. Borzoi (5%), Windsprite (3%), Rhodesian Ridgeback (2%), and Ibizan Hound (2%) showed a low allele prevalence (frequencies less than 10%). None of the dogs tested in these breeds had the T/T genotype. The SERPINF2 c.605 T variant allele was not detected in any Afghan, Pharaoh Hound, or Silken Windhound dog tested.

The SERPINF2 c.605 T variant allele was found in other (non-sighthound) dog breeds, although none of those tested showed an allele prevalence over 20% or a T/T genotype prevalence over 4%. Several dog breeds showed a moderate variant allele prevalence (10 to 20%), including Shetland Sheepdog (19%), Newfoundland (19%), English Bulldog (18%), French Bulldog (12%), Anatolian Shepherd (10%), Cane Corso (10%). T/T genotype prevalence in these breeds ranged from 1 to 4%. Of the remaining 41 dog breeds evaluated, 20 breeds displayed a low variant allele prevalence (1 to 6%), while the variant allele was not detected in 22 breeds. Mixed-breed dogs showed a relatively low variant allele prevalence (5%).

Interestingly the SERPINF2 c.605 T allele was also detected in gray wolf populations. Out of 27 animals surveyed, 3 were heterozygous C/T and 2 were homozygous T/T resulting in a variant allele prevalence of 13%.

Computational analysis by Polyphen-2 (PMID: 20354512) predicted that the p.Ala202Val substitution was possibly damaging with a score of 0.569 (sensitivity 0.88; specificity 0.91). PolyPhen-2 scores can range from 0.0 to 1.0 with higher scores considered more disruptive to protein function.