Intravesical Instillation of Dilute Formalin for the Treatment of Severe Hemorrhagic Emphysematous Cystitis in a Diabetic Dog

Introduction

Emphysematous cystitis is a rare complication of urinary tract infection that is characterized by the accumulation of gas within the wall of the urinary bladder. Although the disorder occurs most often in diabetic dogs, it has also been reported in dogs without diabetes.^1–5 The source of the gas is bacteria, usually Escherichia coli (E. coli), that have invaded the wall of the urinary bladder and produced gas as a metabolic by-product of carbohydrate fermentation.⁵ The condition occurs with a higher incidence in diabetic dogs, because they are at an increased risk for urinary tract infection and suffer from chronic glucosuria.⁶ Emphysematous cystitis typically resolves with appropriate treatment of the urinary tract infection and correction of the underlying glucosuria.⁶ This case report details a case of emphysematous cystitis in a diabetic dog with a chronic, antibiotic-resistant, E. coli urinary tract infection that presented as life-threatening bladder hemorrhage. The dog was treated with intravesical infusion of 1% formalin solution that resolved the hemorrhage and produced a temporary amelioration of the urinary tract infection.

Case Report

A 7-year-old, 4.1-kg, spayed female Yorkshire terrier with a history of poorly controlled diabetes mellitus, bacterial cystitis, and chronic fibrosing hepatitis for approximately 1 year, presented with a 2-day history of lethargy and a 1-day history of severe hematuria. The dog was receiving ursodiol^a (75 mg, per os [PO], sid), lente insulin^b (4 U, subcutaneously [SC], bid), and furadantine^c (15 mg, PO, tid). The dog’s condition had been stable when examined 1 week earlier during a scheduled recheck appointment. At that time, a complete blood cell count (CBC), serum biochemical profile, urinalysis, and urine culture had been performed. The results disclosed significant abnormalities in the blood glucose, alkaline phosphatase, albumin, and total bilirubin [see Table]. These abnormalities were consistent with the dog’s history of diabetes mellitus and hepatic disease and were similar to previous serum biochemical results obtained for the dog. The CBC showed mild anemia, neutrophilic leukocytosis with a slight increase in band neutrophils, mild eosinophilia, and marked monocytosis. A urine sample obtained by cystocentesis contained glucose, blood, bilirubin, and protein. Microscopic examination showed red blood cells (RBCs), white blood cells (WBCs), numerous epithelial cells, and a moderate number of rod-shaped bacteria. Cultured urine grew nonhemolytic E. coli (>100,000 colony-forming units/mL [CFU/mL]) and Enterococcus faecalis (>100,000 CFU/mL). The E. coli was sensitive only to amikacin, showed intermediate resistance to ticarcillin/clavulanic acid and nitrofurazone, and was resistant to all other antibiotics tested, including enrofloxacin and trimethoprim/sulphadiazine. When cultured 1 month earlier, the E. coli organism had been sensitive to nitrofurazone, which had prompted treatment with Furadantin.^c

At the time of current presentation, the dog appeared lethargic and uncomfortable when examined. Physical examination revealed tachycardia (150 beats per minute [bpm]), tachypnea (30 breaths per minute), pale mucous membranes, hypothermia (rectal temperature of 98.8°F), and discomfort on palpation of the caudal abdomen. The dog continually passed small volumes of bloody urine that contained clots, and dysuria was noted. The towel that the owner had used to wrap the dog was heavily stained with bloody urine.

A CBC revealed marked anemia and a moderate leukocytosis with a mature neutrophilia; the remainder of the leukogram was similar to the one obtained 1 week earlier [see Table]. Serum biochemical profile was consistent with previous results and showed decreased albumin and an elevated alkaline phosphatase activity and total bilirubin level [see Table]. In addition, moderate hypernatremia and hyper-chloremia and mild hypokalemia were now present. The results of a coagulation panel [see Table] were within the reference range. Urine obtained by cystocentesis appeared to be frank blood. Urine sediment examination was not performed, but a sample was submitted for bacterial culture.

On abdominal radiographs, the urinary bladder appeared mottled due to the presence of intramural and intraluminal gas [see Figure]. There was a generalized decrease in serosal detail throughout the abdomen. The radiographic findings were compatible with a diagnosis of emphysematous cystitis. The loss of serosal detail had been noted on previous radiographs.

Initial fluid therapy with lactated Ringer’s solution (LRS) supplemented with 16 mEq/L of potassium chloride was given via a jugular catheter at 18 mL per hour. Antibiotic therapy was initiated with a broad-spectrum combination of amikacin^d (15 mg/kg body weight, intravenously [IV], sid) and cefazolin^e (20 mg/kg body weight, IV, tid). These drugs were chosen empirically on the basis of urine culture and sensitivity results obtained 1 week earlier. Hemorrhage from the urinary bladder continued unabated, and it appeared unlikely that antibiotic therapy would act quickly enough to resolve the cystitis. It was judged that more aggressive measures were required to prevent further blood loss. Anesthesia was induced using propofol (3.5 mg/kg body weight, IV to effect), and the dog was intubated with a 5.0-mm endotracheal tube. Anesthesia was maintained with isoflurane and oxygen. A 6-French Foley catheter was placed into the bladder, and the dog was placed in sternal recumbency. The cranial portions of her body were elevated above the caudal portions. The bladder was gently flushed several times using 20 to 30 mL of 0.9% sodium chloride (NaCl). Next, 25 mL of a 1% (volume/volume) formalin solution^f (in 0.9% NaCl) was slowly infused into the urinary bladder. The formalin solution was allowed to stay in the urinary bladder for 10 minutes. The formalin solution was then removed, and the bladder was again flushed with 0.9% saline. The formalin was instilled a second time using the same method. After the formalin treatment was completed, the bladder was completely drained and flushed three times using 20 mL of 0.9% saline each time. The Foley catheter was fixed in place and attached to a closed urine collection system. An epidural injection of morphine^g (0.4 mg) and lidocaine^h (6 mg) was given. This was done because of pain previously noted on physical examination and in anticipation that the formalin treatment may cause additional discomfort. A reduction in the amount of hemorrhage was evident almost immediately after the formalin treatment was completed. Following the procedure, the dog recovered uneventfully and appeared comfortable. She remained alert and later ate a small amount of food. Her heart rate remained around 100 to 120 bpm. Pain control was continued through the night using hydromorphoneⁱ (0.25 mg/kg body weight, IV) when needed. The dog produced bloody urine for several hours following the procedure. Twelve hours later, the hematocrit (Hct) was 15%, and the total protein was 5.4 g/dL. The reduction in the Hct was thought to be due primarily to fluid redistribution into the vascular space rather than ongoing hemorrhage into the urinary bladder. A transfusion of packed RBCs (40 mL at 5 mL/kg per hour) was given, after which the Hct was 20%.

Antibiotic therapy was continued using amikacin^d and cefazolin.^e The urine appeared grossly normal within 24 hours after the formalin infusion. Serum biochemical analyses demonstrated near complete resolution of the previously noted electrolyte abnormalities; mild persistent hypokalemia (3.6 mmol/L) was the only abnormality noted. The dog continued to receive insulin but was also placed on a 5% dextrose solution (in LRS) because the blood glucose was consistently <150 mg/dL when measured. Culture of the urine obtained at the time of emergency admission again yielded a nonhemolytic E. coli. Sensitivity testing was performed using an extended panel of antibiotics. The E. coli isolate was found to be completely resistant to nitrofurantoin but sensitive to amikacin, ceftriaxone, and imipenem.

Steady clinical improvement continued over the next 72 hours. Repeat radiographs of the abdomen demonstrated complete resolution of the emphysematous cystitis detected 3 days earlier.

The dog was discharged from the hospital after 5 days. Treatments consisted of ceftriaxone^j (25 mg/kg body weight, SC bid), lente insulin^b (2 U, SC bid), and ursodiol^a (75 mg, PO, sid). At a recheck examination 5 days later, the owner reported marked improvement in the dog’s condition and resolution of clinical signs of cystitis. Sediment analysis of urine obtained by cystocentesis showed only 2 RBCs/hpf and an occasional WBC/hpf. The dog’s Hct had increased to 27%. One week after ceftriaxone was initiated, the clinical signs had resolved, and urinalysis was unremarkable and bacteria were not seen. Repeat abdominal radiographs showed no signs of emphysematous cystitis. Throughout the course of ceftriaxone treatment, the dog remained free of clinical signs of urinary tract infection. However, despite an 8-week course of ceftriaxone, the dog again developed signs of urinary tract infection, and E. coli (>100,000 CFU) was again cultured from the urine. Unfortunately, the dog died suddenly at home while receiving a second 8-week course of ceftriaxone. A necropsy examination was not performed.

Discussion

Emphysematous cystitis is an uncommon disorder in dogs, cats, and humans.³ Diagnosis requires demonstration of gas in the bladder lumen or within the bladder wall. In dogs and cats, emphysematous cystitis is most commonly diagnosed in patients with glycosuria due to diabetes mellitus, but it has been reported rarely in patients with other causes of glycosuria and in some without glycosuria.^1–4 In theory, infection by any gas-producing bacteria could result in the condition. In diabetic patients, appropriate therapy with insulin and antibiotics generally resolves the infection and cystic emphysema; insulin treatment facilitates resolution by diminishing glucosuria, which is a substrate for bacterial gas production.⁶

In this case report, the clinical presentation was one of urinary bladder hemorrhage, requiring immediate and aggressive treatment. It was believed that routine treatment with antibiotics alone would not be expedient, because this dog suffered from a chronic, antibiotic-resistant, bacterial cystitis. Intravesical instillation of a dilute formalin solution has been used to treat cyclophosphamide-induced hemorrhagic cystitis in dogs and humans.^7–10 It has not been reported as a treatment for hemorrhagic emphysematous cystitis.

Formalin exerts its in vivo effect by the same mechanism as its fixative action on tissue specimens, through coagulation of the superficial mucosal tissue and protein hydrolysis.¹⁰ Initial treatment recommendations for hemorrhagic cystitis in human patients used 10% or 4% formalin concentrations, which produced severe adverse effects in some patients, including incontinence and papillary necrosis.¹¹ Reported minor side effects included transient obstructive uropathy, pain, discomfort, and a reduction in bladder capacity.¹⁰ Subsequently, it was demonstrated that a 1% formalin solution was equally as effective as 10% or 4% solutions.⁹ The only reported adverse effect of 1% formalin treatment for cyclophosphamide-induced cystitis in two dogs was intermittent urinary incontinence in one of the dogs.⁷

No complications were observed in the patient of this study. Pain was managed proactively in this patient, and an indwelling urinary catheter was maintained for 48 hours following the procedure. These measures probably served to minimize side effects, and there were no indications of discomfort as a result of the procedure. Blood loss was drastically reduced within 24 hours, and radiographic resolution of bladder abnormalities was complete by 72 hours posttreatment.

In the patient of this study, hemorrhage resolved after a single formalin treatment. Many human patients treated with formalin demonstrate dramatic improvement after a single treatment, although multiple treatments are necessary in some patients.⁸ In the two dogs with cyclophosphamide-induced cystitis treated with 1% formalin, a single treatment resolved the problem.⁷ Despite the rapid improvement after treatment, the authors do not believe that formalin treatment alone was responsible for the rapid resolution of this dog’s cystitis, as more established therapy was also provided. In this situation, it was not possible to directly examine any role formalin might have had in resolving the emphysema and in decreasing the bacterial population in the bladder. Thus, it remains possible that the rapid improvement observed after formalin infusion could have been due to a direct action of killing the bacteria. Dilute formalin infusion does not supplant the use of established therapies for emphysematous cystitis, and the authors do not recommend the routine use of this therapy. However, formalin therapy may have specific indications for the treatment of life-threatening hemorrhagic cystitis, including that resulting from emphysematous cystitis, as in this dog, when it is unlikely that conventional therapy will act quickly enough to be beneficial.