Severe Hematometra in a Dog With Cystic Endometrial Hyperplasia/Pyometra Complex

Case Report

An 18-month-old, 27.3-kg, nulliparous, intact female German shepherd dog was referred for evaluation of severe hemorrhagic vulvar discharge. The discharge began 6 days prior to presentation and initially was mild and serosanguineous. The bloody discharge slightly increased in volume the night prior to presentation and then turned into frank vulvar hemorrhage by the morning. The dog’s last estrus occurred 4 months prior to admission and was reported by the owner to be a normal heat cycle, as was the previous heat cycle. The owner believed the dog to be in estrus when the hemorrhagic discharge began. The owner stated that there was no possibility that the dog had been bred, as she was housed in a run with 6-foot fencing at all times while outdoors. In addition to the vulvar discharge, the dog had one episode of dark, tarry stool the night before presentation. The dog was vaccinated 2 months prior to presentation and had not been given any medications during that time. No anticoagulant rodenticides were kept within the house or in the yard. There was no history of tick infestation noted by the owner.

On physical examination, the dog was quiet, alert, responsive, and approximately 5% to 7% dehydrated. The dog was afebrile, had a normal heart rate, and was tachypneic (respiratory rate, 80 breaths per minute); however, all lung fields were normal on auscultation. Examination of the skin and mucous membranes did not reveal ecchymoses or petechiae. The dog exhibited mild discomfort on abdominal palpation, but no other palpable abnormalities were detected. Digital rectal examination revealed melena. Profuse hemorrhagic discharge was noted from the vulva, and the dog occasionally passed long, thread-like blood clots through the vulva. Digital vaginal examination revealed no other abnormalities.

Initial diagnostic testing consisted of an activated clotting time (ACT), packed cell volume (PCV), total protein (TP), blood urea nitrogen (BUN), blood glucose, peripheral blood smear, and abdominal radiographs. Additionally, whole blood was submitted to a reference laboratory for complete blood count (CBC), reticulocyte count, platelet count, complete serum biochemical analysis, coagulation profile, and rickettsial serology. At presentation, the PCV was 31% (reference range, 37% to 55%) and the TP was 5.0 mg/dL (reference range, 5.4 to 7.5 mg/dL), consistent with a recent hemorrhagic episode. The BUN, blood glucose, and ACT (60 seconds; reference range, <120 seconds) were all within reference ranges. The blood smear demonstrated mild to moderate anisocytosis and polychromasia and severe thrombocytopenia. Fewer than 20 platelets were noted on the entire peripheral blood smear. Abdominal radiographs were within acceptable limits, except for the presence of radiopaque material (consistent with rocks) within the stomach. Voided urine was clear and yellow, suggesting that the source of hemorrhage was the reproductive tract rather than the urinary tract.

Based on the historical information obtained, physical examination findings, and the initial diagnostic tests listed above, a presumptive diagnosis of immune-mediated thrombocytopenia (IMTP) was made. The melena was believed to be due to ingestion of blood; however, gastrointestinal hemorrhage was possible. The dog was placed on prednisolone acetate (2.2 mg/kg body weight, intramuscularly [IM], divided bid), lactated Ringer’s solution (160 mL per hour, intravenously [IV]), and famotidine (1 mg/kg body weight, IV, sid), pending the laboratory results and an abdominal ultrasound scheduled on day 2. Vitamin K₁ therapy for anticoagulant rodenticide toxicity was not initiated at that time because of a low index of suspicion and the presence of a normal ACT.

During the course of the first night, the dog became increasingly lethargic, tachycardic, and tachypneic. A recheck PCV was 19% 6 hours postadmission. Because of the acute decline in PCV and patient deterioration, the dog was given a 450-mL packed red blood cell (pRBC) transfusion over 2 hours. The following morning (day 2), the PCV had increased to 27% and there was an improvement in clinical signs. The hemorrhagic discharge, however, continued unabated throughout the night.

Laboratory tests from day 1 showed the following abnormalities: mild normocytic, normochromic, nonregenerative anemia (PCV, 31%; reticulocyte count, <1%); hypoproteinemia (4.7 g/dL; reference range, 5.4 to 7.5 g/dL); hypoalbuminemia (2.2 g/dL; reference range, 2.7 to 4.4 g/dL); high-normal BUN (25 mg/dL; reference range, 6 to 25 mg/dL); hyperphosphatemia (7.0 mg/dL; reference range, 2.5 to 6.0 mg/dL); and hypokalemia (3.2 mEq/L; reference range, 3.6 to 5.5 mEq/L). The platelet count showed mild thrombocytopenia (160 × 10³/μL; reference range, 170 to 400 × 10³/μL); however, platelet clumping was noted, preventing a precise determination of platelet number. Mild hyperfibrinogenemia was also noted (453 mg/dL; reference range, 150 to 400 mg/dL). The prothrombin time (PT), partial thromboplastin time (PTT), and D-dimer were within normal limits. The rickettsial serology was negative for Ehrlichia canis, Rickettsia ricketsii, and Borrelia burgdorferi.

As a result of these laboratory findings, IMTP was no longer considered, and prednisolone therapy was discontinued after the first dose. Based on the normal PT, PTT, and D-dimer, a coagulopathy and disseminated intravascular coagulopathy (DIC) were less likely.

On day 2, an abdominal ultrasound demonstrated mildly dilated, fluid-filled uterine horns. The uterine wall had a cystic, honeycomb appearance. The uterine body and cranial vaginal lumen were fluid-filled and contained hyper-echoic, polypoid masses [Figure 1]. The ovaries were not identified. The liver was diffusely, mildly hypoechoic. All other ultrasonographic findings were within normal limits.

Based on the results of abdominal ultrasound, an exploratory laparotomy was recommended. Prior to the procedure, the dog’s PCV dropped to 17%, and a 225-mL pRBC transfusion was initiated. Intravenous, broad-spectrum antibiotic therapy (cefazolin, 22 mg/kg body weight, IV, tid) was instituted preoperatively. The laparotomy revealed a slightly enlarged, fluid-filled uterus with palpable, intraluminal and intramural irregularities. The uterine wall felt thickened and edematous. An ovariohysterectomy (OHE) was performed. In addition, a liver biopsy was performed to further investigate the decreased echogenicity seen during ultrasonography. There was no significant intra-or postoperative hemorrhage.

Postoperatively, the uterine body was incised, revealing a diffuse, cystic appearance to the endometrium with a large amount of blood and blood clots within the lumen [Figure 2]. Additionally, there were several focal accumulations of thick, yellow material presumed to be purulent exudate; however, cytology or bacterial culture was not performed to confirm this suspicion. The dog recovered uneventfully from surgery and was maintained on IV fluids (lactated Ringer’s solution, 100 mL per hour) and antibiotics (cephalexin, 22 mg/kg body weight, per os [PO], tid for 7 days). The dog was discharged 3 days later without further hemorrhage.

Histopathology confirmed the presence of marked hyperplasia of the endometrial glands that were cystically dilated and filled with a proteinaceous, secretory material mixed with blood and neutrophils. The ovaries were normal. The final histopathological diagnosis was severe pyometra with marked cystic endometrial hyperplasia (CEH). The liver biopsy revealed mild, hydropic, vacuolar hepatopathy that was likely secondary to nonhepatic disease.

Discussion

Cystic endometrial hyperplasia is a progressive and pathological endometrial thickening caused by an increase in the size and number of endometrial glands. These glands may possess secretory material and actively secrete a thin or viscous fluid into the uterine lumen, leading to hydrometra or mucometra.¹² Chronic or repeated progesterone stimulation, or both, are thought to lead to the exaggerated endometrial response characteristic of CEH. Uncomplicated CEH is not commonly diagnosed because of the lack of clinical signs and the need for uterine biopsy. There is no therapy available for uncomplicated CEH.¹

Pyometra is typically considered to be a disease of intact, middle-aged bitches, but it can occur in young dogs. It is a progesterone-mediated disorder of diestrus and is defined as the accumulation of purulent exudate within the uterine lumen secondary to intrauterine bacterial infection. The most commonly isolated organism is Escherichia coli (E. coli), a member of the normal vaginal flora. Additional isolated bacteria include Staphylococcus, Streptococcus, Pseudomonas, Proteus, Pasteurella, and others.³ Mild to severe bacteremia and endotoxemia are common sequelae.¹² Cystic endometrial hyperplasia typically precedes pyometra; however, pyometra can occur without CEH. Dogs <6 years of age are less likely to have CEH prior to the onset of pyometra.¹

Regardless of whether CEH is present, pyometra almost always begins during diestrus while the uterus is under the influence of progesterone. Occasionally, a bitch will develop pyometra during diestrus but will fail to show clinical signs until anestrus.¹ Progesterone concentrations are low (<1.0 ng/mL) during anestrus and proestrus, and then they begin to rise during estrus, with concentrations increasing to >2.0 ng/mL during estrus. Progesterone concentrations continue to increase through the first few weeks of diestrus before reaching a plateau, followed by a slow decline as the bitch approaches anestrus. In the normal bitch, progesterone concentrations are increased (>2 ng/mL) for the first 9 to 12 weeks of diestrus and can occasionally be >40 ng/mL.¹ Progesterone suppresses myometrial contractions while stimulating endometrial growth and glandular secretions.¹ This allows accumulation of vaginal secretions, providing an optimal environment for bacterial growth. Additionally, leukocyte inhibition in the progesterone-primed uterus supports further bacterial growth.¹

The most common clinical complaint is serosanguineous to mucopurulent vaginal discharge, often noticed 4 to 8 weeks after standing heat.¹ Other clinical signs frequently reported include lethargy and depression; inappetence or anorexia; polyuria, polydipsia, or both; and vomiting.¹ Physical examination findings consistent with pyometra include dehydration, depression, fever, and vaginal discharge.¹² Vaginal discharge is not present with a closed-cervix pyometra. Uterine enlargement may be noted on abdominal palpation or during rectal examination.

Routine blood work aids in the diagnosis, but abnormalities are not definitive for pyometra or are nonspecific. Complete blood count results can vary, often showing an absolute neutrophilia with variable degrees of left shift. Presence of an absolute leukocytosis is also variable. A mild normocytic, normochromic, nonregenerative anemia often develops secondary to chronic disease and generally resolves following treatment. Hyperproteinemia and hyperglobulinemia are possible secondary to dehydration or antigenic stimulation. Prerenal azotemia is likely if dehydration is present. Mild to moderately elevated liver enzymes (i.e., alanine transferase and alkaline phosphatase) are possible due to hepatocellular damage secondary to bacteremia/endotoxemia, hepatocellular hypoxia secondary to decreased hepatic circulation with dehydration, or both.¹²

Cystocentesis is not recommended if pyometra is suspected; however, a midstream free-catch urine sample can be collected for urinalysis. Results of urinalysis are variable. Urine specific gravity is often normal early in the course of the disease; however, isosthenuria is possible secondary to polyuria with compensatory polydipsia.¹ Polyuria is thought to be due to reversible renal tubular damage by toxins produced by E. coli. This leads to decreased activity of antidiuretic hormone, producing a secondary nephrogenic diabetes insipidus.¹ Other mechanisms have been proposed as causes of isosthenuria, such as renal tubular immune complex injury and renal medullary solute washout.¹

Diagnostic imaging generally confirms a presumptive diagnosis of pyometra. The classic radiographic finding is a coiled, tubular structure of soft tissue or fluid opacity in the caudal abdomen between the descending colon and urinary bladder. Loops of bowel are often displaced craniodorsally and medially.⁴ Ultrasonography is a valuable diagnostic tool, allowing the measurement of uterine wall thickness and demonstration of fluid within the uterine lumen. Occasionally, cystic changes consistent with CEH may be seen in the uterine wall.⁴

The preferred treatment for pyometra is OHE.¹² Severely ill bitches should be stabilized first with IV fluids and broad-spectrum antibiotics (e.g., amoxicillin/clavulanic acid, cephalosporins, or trimethoprim-sulfamethoxazole) that are effective against E. coli and other common bacterial organisms listed above.¹ Ovariohysterectomy should be performed as soon as possible due to the life-threatening complications associated with bacteremia and endotoxemia. While surgery is the treatment of choice, medical therapy with prostaglandin F_2α (PGF_2α) and broad-spectrum antibiotics to preserve the reproductive potential of the bitch has been reported.¹²⁵⁶ The goal of using PGF_2α is to cause the physical expulsion of vaginal contents through the combined mechanism of myometrial contraction and cervical relaxation.² Medical management of closed-cervix pyometra has been reported;¹⁶ however, bitches with closed-cervix pyometra may be so ill that immediate OHE is recommended.¹

In the case reported here, a diagnosis of pyometra was not made until exploratory laparotomy. Cystic uterine changes observed after OHE were later confirmed by histopathology to be a severe case of CEH/pyometra complex. Many of the common clinical findings associated with pyometra were not present in this case. The clinical history was not overtly suggestive of pyometra, as the vaginal discharge began 4 months after the previous heat cycle. However, it is possible that the owner misjudged the date of the previous estrus. Physical examination findings were also inconclusive, as the dog was afebrile; quiet, but not lethargic or depressed; had frank hemorrhagic vaginal discharge, rather than serosanguineous or mucopurulent discharge; and was not polyuric or polydipsic. An early diagnosis of pyometra was also confounded by an errant platelet estimate on the peripheral blood smear. Abdominal radiographs failed to demonstrate a fluid-filled uterus in the caudal abdomen.

Hematometra (i.e., collection of blood within the uterine lumen) and metrorrhagia (i.e., uterine hemorrhage between heat periods) are uncommon clinical presentations. Reported etiologies include postpartum subinvolution of placental sites,⁷⁸ anticoagulant rodenticide toxicity⁹ and other acquired coagulation deficiencies, uterine trauma, neoplasia,¹⁰¹¹ placental necrosis, congenital coagulation deficiency,¹² idiopathic prepubertal metrorrhagia,¹³ postpartum endometritis,¹ and uterine serosal inclusion cysts.¹⁴ The differential diagnoses initially considered in this case included idiopathic thrombocytopenic purpura, other platelet disorders (e.g., thrombocytopenia secondary to rickettsial diseases), anticoagulant rodenticide toxicity, DIC, and congenital coagulation disorders. Congenital coagulation disorders seemed unlikely based on a lack of prior bleeding episodes. Anticoagulant rodenticide toxicity was also considered less likely because of a lack of known exposure and the severe thrombocytopenia noted on the initial peripheral blood smear. Mild to severe thrombocytopenia has been reported secondary to rodenticide toxicities; however, the degree of anemia did not correlate with the degree of thrombocytopenia initially thought to be present in this case. Given the 6-day history of hemorrhagic discharge, a prolonged ACT, PT, and PTT would have been likely. Acquired coagulopathies, including DIC, were eventually ruled out based on the normal PT, PTT, and D-dimer.

Based on the clinical history, physical examination, and peripheral blood smear, IMTP was initially considered the primary differential diagnosis. Additional causes for throm-bocytopenia that were considered included decreased platelet production (e.g., immune-mediated megakaryocytic hypoplasia, myelopthisis, and cyclic thrombocytopenia) and increased platelet destruction, sequestration, or utilization (e.g., vaccine or drug-induced thrombocytopenia, DIC, vasculitis, splenomegaly, endotoxemia, and neoplasia). Corticosteroid therapy was initiated based on a presumptive diagnosis of IMTP. A platelet count of approximately 25 × 10³ platelets/μL or less is required to induce spontaneous hemorrhage.¹⁵ An automated platelet count performed by a reference laboratory, however, demonstrated only mild thrombocytopenia with platelet clumping. As a result, corticosteroid therapy was immediately discontinued. One would also expect petechiae or ecchymoses to be present with severe thrombocytopenia; however, petechiae were not noted during physical examination. Possible causes for errant platelet counts obtained from peripheral blood smears include platelet clumping, clustering at the feathered edge, or both, and inadequate staining. No platelet clumps or clustering at the feathered edge were noted on the slide. In this case, the most likely cause for the errant platelet estimate was inadequate staining.

Additional diagnostics that could have been performed in this case include vaginal cytology, vaginoscopy, buccal mucosal bleeding time (BMBT), thoracic radiographs, and von Willebrand’s factor assay. Vaginal cytology would have been useful in determining the stage of the estrous cycle. Cytology may have also demonstrated the presence of toxic or degenerative neutrophils or bacteria or both, thus increasing the suspicion for pyometra prior to abdominal ultrasound. Vaginoscopy was not available as a diagnostic tool at the time of this case. A BMBT was not performed initially due to thrombocytopenia detected on the in-house blood smear; however, platelet dysfunction could have been ruled out by a BMBT after a normal platelet count was obtained from the reference laboratory. Thoracic radiographs were not taken because of a lack of clinical signs referable to cardiac or respiratory disease.

German shepherd dogs are predisposed to von Wille-brand’s disease. Mucosal bleeding is common, often occurring at the gingiva, nose, and gastrointestinal or urogenital tract.¹⁶ It is possible that a deficiency in von Willebrand’s factor in this dog may have been a complicating feature. A von Willebrand’s factor assay was not performed in this case due to lack of prior bleeding episodes as well as a lack of intra- and postoperative hemorrhage.

Ultrasonography finally aided in identifying intraluminal uterine fluid and cystic changes in the uterine wall, leading to the decision to perform an exploratory laparotomy. Histopathology revealed CEH/pyometra and ruled out other causes of hematometra.

A progesterone assay on the original serum submitted to the reference laboratory was performed after the histopathological diagnosis of CEH/pyometra complex was obtained. The progesterone level was 0.2 ng/mL, consistent with anestrus or proestrus. In this case, the bitch likely had pyometra for some time before displaying clinical signs.

Conclusion

Pyometra should be suspected in any ill bitch in the diestrus phase of the luteal cycle that is displaying clinical signs of fever, anorexia, depression, and polyuria/polydipsia, with or without vaginal discharge. Although the strict definition of pyometra is accumulation of purulent exudate within the uterine lumen, it is important to remember that serosanguineous discharge is common. To the best of the authors’ knowledge, this is the first reported case of frank uterine hemorrhage occurring with CEH/pyometra complex in the bitch. Typically, a diagnosis of pyometra can be reached by combining historical, physical examination, clinicopathological, radiographic, and, if available, ultrasonographic findings. This case demonstrates that pyometra remains a challenging diagnosis.