Candida peritonitis is reported in people and is associated with significant morbidity and mortality compared with sterile or bacterial peritonitis. Recognized predisposing risk factors include peritoneal dialysis, hollow viscous organ perforation, abdominal surgery, inflamed intestinal mucosa, antimicrobial administration, and immunosuppression. In this report, we describe 5 cases of dogs with peritonitis complicated by Candida spp; 3 dogs with C albicans, one dog with C albicans and C glabrata, and one dog with C glabrata only. The 3 dogs with C albicans peritonitis presented with duodenal perforation due to NSAID therapy, intestinal resection and anastomosis following postspay-surgery dehiscence, and intestinal foreign body removal. The 2 dogs with C glabrata peritonitis had undergone cholecystectomy due to gall bladder rupture and dehiscence of intestinal biopsy removal sites following exploratory laparatomy. In all cases, initial diagnosis of fungal peritonitis was made via cytologic examination of peritoneal effusions, which revealed marked pyogranulomatous inflammation with numerous 3–8 μm oval, deeply basophilic yeast organisms with thin clear capsules noted within phagocytes and extracellularly. In addition, germ tube formation, hyphae, and pseudohyphae were rarely seen in some of the cases with pure C albicans. Identity of the organisms was determined by culture in all cases and confirmed by PCR in 3 cases. Candida spp. are commensals normally inhabiting the alimentary, the upper respiratory, and the lower urogenital tracts of mammals. They are opportunistic pathogens that can invade and colonize tissue when a patient is immune-compromised or there is disruption of the mucosal barrier. Candida peritonitis should be considered in patients with peritoneal contamination with gastrointestinal or biliary contents.
Candida peritonitis is reported frequently in people and is associated with significantly increased mortality rates compared with either sterile peritonitis or intra-abdominal sepsis due to bacteria alone.[1-3] Candida spp. are the most common cause of invasive mycoses, including fungal peritonitis in people.[2, 4] In contrast to the numerous publications in human medicine, there is a dearth of information on Candida peritonitis in the veterinary literature. Only 2 previous cases involving dogs have been reported, one following enterotomy for removal of an intestinal foreign body, and the other one subsequent to intestinal resection and anastomosis for removal of a jejunal mass.[5, 6]
Candida spp. are normal inhabitants of the gastrointestinal (GI) and biliary tracts of people and animals; therefore, peritonitis may develop secondary to overgrowth and extension from these sites in patients with a compromised mucosal barrier. In people, Candida peritonitis is seen following peritoneal dialysis, GI surgery with spillage of enteric contents into the peritoneal space, GI ulceration, hollow viscous organ perforation, and mucosal neoplasia. Other less common inciting causes of Candida peritonitis in people include pancreatitis, appendicitis, diverticulitis, diabetes mellitus, and penetrating trauma. Most notably, the majority of cases of Candida peritonitis were receiving systemic, broad-spectrum antimicrobial therapy at the time of diagnosis.[1-4, 7-17]
We have recognized 5 cases of Candida peritonitis from December 2008 to April 2011, all diagnosed initially on cytology of peritoneal fluid and subsequently confirmed by culture, or culture and PCR in 3 of the cases. The purpose of this study was to describe the clinical parameters associated with septic peritonitis complicated by Candida spp. in dogs, a condition which is potentially under-diagnosed, under-reported, or possibly increasing in frequency as noted with people.
An 11-year-old intact male Poodle was presented to Oklahoma State University, Center for Veterinary Health Sciences (OSU-CVHS), with a history of vomiting, abdominal distention, increased serum activities of alkaline phosphatase, γ-glutamyl transferase, and hyperbilirubinemia. The dog had received amoxicillin and clindamycin (dosage and administration route unknown) prescribed by the referring veterinarian.
An abdominal ultrasonographic examination revealed a moderate amount of peritoneal fluid, widespread vascular mineralization, and a markedly thickened and irregular gallbladder wall. The abdominal fluid had a total nucleated cell count (TNCC) of 51,000/μL (CELL-DYN 3500 analyzer; Abbott Diagnostics, Abbott Park, IL, USA), and a total protein of 5.6 g/dL determined by refractometry. Cytologic examination of direct smears of the abdominal fluid stained with an aqueous Romanowsky stain (Hematek 2000; Siemens Healthcare Diagnostics, Deerfield, IL, USA) demonstrated marked pyogranulomatous inflammation with abundant golden-to-dark green pigment, consistent with bile, present extracellularly and within macrophages. Cytologic diagnosis was bile peritonitis. No infectious organisms were identified, and subsequent aerobic and anaerobic bacterial cultures were negative.
The dog was started on intravenous ampicillin/sulbactam and enrofloxacin. Exploratory abdominal laparotomy revealed a ruptured gall bladder and a cholecystectomy and duodenotomy were performed. Following surgery, the effusion persisted, despite antimicrobial therapy, and the dog's condition began to deteriorate. Abdominal fluid collected 6 days postsurgery had a TNCC of 96,600/μL, and a total protein of 5.3 g/dL. Cytologic examination demonstrated marked pyogranulomatous inflammation with numerous yeast organisms present extracellularly as well as within neutrophils and macrophages (Figure 1A). The basophilic organisms were round to oval, approximately 3–8 μm in diameter, and had a clear, thin circumferential halo. Few organisms showed narrow-based budding. Rare elongated yeast organisms consistent with germ tube formation were noted phagocytized by macrophages (Figure 1B). No bacteria were seen. A second exploratory laparotomy was performed and revealed an abscess at the gallbladder fossa, which was debrided and omentalized. The dog subsequently developed anuric renal failure and was euthanized 2 days after the second surgery. While septicemia was suspected, blood cultures were not performed and necropsy was declined by the owner. No antifungal therapy was initiated in this dog prior to euthanasia.
Culture of the second abdominal fluid sample (sheep blood agar, MacConkey agar, and Sabouraud dextrose agar) revealed the presence of E coli and 2 morphologically distinct fungal colonies consistent with Candida spp. Colonies were identified as Candida albicans and C glabrata, respectively (MicroScan, Rapid Yeast Identification, Siemens Healthcare Diagnostics Inc., West Sacramento, CA, USA). Histopathology of the gallbladder showed severe, diffuse, chronic-active, necrotizing suppurative cholecystitis with fibroplasia. No infectious organisms were identified in the biopsy samples using Periodic acid-Schiff (PAS) stain or Gomori's methamine silver (GMS).
An 8-month-old spayed female miniature Australian Shepherd had a history of dehiscence of an ovario-hysterectomy incision performed by the referring veterinarian that resulted in exteriorization of the intestines. The dog had undergone 2 subsequent exploratory laparatomies at the primary veterinary hospital with resection of compromised bowel, and received intravenous and oral antibiotics, including cephalexin, enrofloxacin, cefazolin, and metronidazole. The dog was then presented to OSU-CVHS for continued lethargy, anorexia, and mild-to-moderate abdominal distention.
Ultrasound-guided percutaneous peritoneal aspiration yielded a moderate amount of yellow, flocculent fluid. Due to the nature of the fluid, cell counts were not obtained. Cytologic examination of aqueous Romanowsky-stained direct smears demonstrated pyogranulomatous inflammation with a few fungal yeast present extracellularly and within neutrophils and macrophages. The yeast organisms were round to oval, basophilic, approximately 3–8 μm in diameter, and had a clear, thin halo. Rare yeast organisms had a protrusion with parallel sides representing germ tube formation.
Culture of the fluid revealed Enterococcus faecium and a fungal organism, which was identified as C albicans, confirmed by PCR amplification and sequencing of the entire internal transcribed spacer (ITS) region of the ribosomal RNA gene, comprising ITS1, 5.8S, and ITS2, followed by BLAST analysis of the gene with GenBank sequences (PCR was performed at Washington Animal Disease Diagnostic Laboratory at Washington State University). The isolate had 100% sequence identity with previously reported sequences of C albicans.
The dog was continued on intravenous amoxicillin/sulbactam, enrofloxacin, and metronidazole and started on fluconazole intravenously. Exploratory laparotomy at OSU-CVHS revealed dehiscence of a previous resection and anastomosis site with areas of necrosis, diffuse intestinal adhesions, free abdominal fluid, and gross contamination with intestinal contents. An intestinal resection and anastomosis with serosal patching was performed. Given the degree of contamination, the abdomen was closed with a modified open technique using a suction tube and continuous suction until the effusion resolved. Histopathology of the excised intestinal tissue obtained at OSU-CVHS was evaluated and severe, chronic-active peritonitis with extensive serosal granulation tissue formation and fibrosis was reported. No fungal organisms were observed in tissue sections, including those stained with PAS and GMS.
The dog's condition drastically improved, and it was discharged with continuing treatment of several weeks with oral trimethoprim-sulfamethoxazole and fluconazole. The dog is alive without evidence of morbidity at follow-up 6 months after discharge.
A 7-year-old intact male English Mastiff was presented to a private referral hospital for vomiting, lethargy, and anorexia. The dog had a history of receiving carprofen for an unknown duration for osteoarthritis treatment and had been started on sulfadimethoxine and ormetoprim one week prior for a suspected urogenital infection. Physical examination revealed a distended abdomen with a palpable fluid wave. Ultrasonographic examination demonstrated a large amount of moderately echogenic fluid in the abdominal cavity. Direct smears of the fluid stained with an aqueous Romanowsky stain were submitted for cytologic evaluation. The slides revealed marked pyogranulomatous inflammation with numerous yeast organisms, which were round to oval, 3–8 μm in diameter, and deeply basophilic with a thin clear halo. Yeast were present extracellularly as well as phagocytized within neutrophils and macrophages. Rare bacterial cocci were also noted within neutrophils.
An exploratory laparotomy was performed, but the dog suffered cardiac arrest during surgery and could not be reanimated. On necropsy, 2 duodenal perforations with about one cm diameter were present adjacent to the bile duct. A large amount of debris and purulent fluid was also evacuated from the abdomen. This was collected for culture via previously described methods and C albicans was isolated; however, no bacterial growth was identified. Tissue samples were not submitted for histopathology. It was suspected that the ulceration occurred as a complication of NSAID administration.
A 4-year-old spayed female American Bulldog was presented to a primary care veterinary hospital for vomiting and lethargy. Exploratory laparotomy revealed a foreign body made of fabric in the stomach, which extended into the proximal jejunum. An enterotomy was performed, and the foreign body removed. The enterotomy and laparotomy sites were closed, and the dog was started on intravenous metronidazole and cephalexin. Over the next few days, the dog started to improve and was switched to oral amoxicillin and clavulanate.
Five days after surgery, the dog became anorectic, lethargic, and started vomiting. Abdominal ultrasound revealed a copious amount of free fluid in the abdomen, which was collected and submitted for cytology and culture. The peritoneal fluid had a TNCC of 47,200/μL and a total protein of 3.5 g/dL. Cytologic examination of direct smears stained with an aqueous Romanowsky stain revealed marked pyogranulomatous inflammation with numerous fungal organisms, which were noted in several forms. Most common were yeast, which were morphologically similar to Candida yeast. The yeast were present within neutrophils and macrophages and also extracellularly. Less commonly, there were yeast with germ tube projections, pseudohyphae with obvious constrictions at septa (Figure 1C) and, rarely, true hyphae with parallel sides lacking constrictions at septal sites (Figure 1D). C albicans was cultured from the fluid, and the identity was confirmed by PCR amplification and sequence of the ITS region of the ribosomal RNA gene as previously described. No bacterial organisms were isolated.
A second exploratory laparotomy was performed to identify the source of the septic effusion. The previous enterotomy incision had dehisced and adhesions had formed over the small intestines. Sections of bowel in the area of the dehiscence were collected for histopathology. The enterotomy incision was repaired, and the abdomen lavaged. The dog was restarted on intravenous enrofloxacin and cephalexin as well as oral fluconazole.
The dog's condition continued to deteriorate further over the next few days, and a third exploratory laparotomy was performed. The intestines had extensive adhesions as well as numerous ulcerations on their serosal surfaces. The enterotomy incision appeared unstable and near dehiscence. The owner elected euthanasia due to the grave prognosis.
Fresh samples from the affected sections of the intestine were submitted for histopathology postmortem. Histopathology confirmed marked, pyogranulomatous enteritis, serositis, and peritonitis. GMS and PAS staining revealed numerous fungal elements near the serosal surface. These were mostly small yeast with moderate numbers of short hyphae and pseudohyphae (Figure 2).
A 2-year-old spayed female Standard Poodle was presented to OSU-CVHS for evaluation of abdominal distension, following collection of intestinal biopsies for chronic small bowel diarrhea associated with weight loss and anorexia. An abdominal exploratory surgery had been performed by the referring veterinarian and no obvious gross abnormalities involving the GI tract were observed. Biopsies were taken from the duodenum, jejunum, ileum, and an enlarged ileocecal lymph node was excised. Histopathology of the lymph node, ileum, and duodenum revealed atypical lymphoid hyperplasia, which consisted of an expanded population of intermediate lymphocytes and an inverted follicle morphology where small lymphocytes are centrally located and surrounded by a mantle of proliferating lymphoblasts. These findings were suspected to be secondary to food allergies or other inflammatory processes, including inflammatory bowel disease. All other tissue sections were unremarkable.
Two days after the surgery by the referring veterinarian, the dog developed abdominal distention and continued to be lethargic and anorexic. The patient was started on intravenous cefazolin and enrofloxacin. A second exploratory laparotomy was performed and revealed dehiscence of one of the jejunal biopsy sites. The dehiscence was repaired, and the dog improved transiently. However, one week after the second exploratory surgery, the dog's anorexia and lethargy relapsed, fluid distention again developed, and the dog was referred to OSU-CVHS.
Upon presentation, the dog was alert with a severely distended abdomen and a palpable fluid wave. Abdominocentesis was performed, and the fluid had a TNCC of 31,100/μL and a total protein of 2.5 g/dL. Cytology revealed suppurative to pyogranulomatous inflammation with occasional intracellular yeast within neutrophils and macrophages. The yeast appeared consistent with Candida spp. Rare bacterial cocci phagocytized by neutrophils were also seen. Cultures of abdominal fluid yielded E coli, Enterococcus faecium, as well as C glabrata. The identity of the fungal isolate was also confirmed with PCR amplification and sequencing of the ITS region of the ribosomal RNA gene.
An abdominal exploratory surgery was performed and 3 L of effusion were collected from the abdomen. Gross debris was present within the abdomen secondary to dehiscence of all previous biopsy sites. The sites were debrided and closed primarily with serosal patching. The abdomen was lavaged and closed with a modified open abdominal drainage system. The patient was started on intravenous fluconazole and enrofloxacin, and oral chloramphenicol. The dog improved following surgery, and the abdominal drain was removed once the effusion resolved. The abdomen was then closed. Repeat culture of the abdominal fluid done via previously described methods prior to closure revealed bacterial infection, including E coli and Pseudomonas aeruginosa. No fungal organisms were cultured. Based on a sensitivity test, the dog was discharged on oral fluconazole, and ticarcillin and clavulanate. At the time of this writing, the dog is alive and doing well with treatment for suspected food allergies.
This report describes 5 dogs with Candida peritonitis, a condition reported only twice previously in the veterinary literature.[5, 6] All dogs in this study had a history of antimicrobial administration coupled with either intestinal and/or biliary surgery (4 cases), or intestinal ulceration associated with administration of a NSAID (one dog). Likewise, the 2 previously reported canine cases were also identified following enteric surgery.[5, 6] These findings are consistent with the predisposing factors reported in people.
Candida spp. are commensals normally inhabiting the alimentary tract as well as the upper respiratory tract and the lower urogenital tract of most mammals.[1, 10] Overgrowth of fungal organisms is physiologically inhibited by normal intestinal flora, intestinal enzymes, peristalsis, and the host's immune system. Antimicrobial use, which allows fungal overgrowth by altering the intestinal microflora, is consistently identified as a risk factor in human studies. These opportunistic pathogens can spread to the peritoneum when there is disruption of the mucosal barrier resulting from conditions such as surgery, perforation, inflammation, or neoplasia.[1-4, 7, 8, 10-17]
Historically, there has been considerable debate in the human literature as to the significance of isolating a GI commensal such as Candida spp. from peritoneal fluid samples in patients with peritonitis. It has been suggested that these organisms may simply be a marker of severe disease without being pathogenic themselves.[19, 20] However, several lines of evidence indicate that these organisms are acting as true pathogens. Many studies of peritonitis in people have shown that patients with positive bacterial culture had a higher mortality rate than those with sterile culture, and that those patients with both bacterial and fungal organisms present in peritoneal fluid had an even higher number of fatalities than those with bacteria alone.[1-3] In people, the association of Candida with cholecystitis similar to case 1 in our study is 100% lethal. Furthermore, while most patients with Candida peritonitis have concurrent bacterial sepsis, in some cases fungal sepsis is present as a sole entity, similar to case 4 in our study, which is associated with morbidity and mortality.[19, 20]
Candida infection is a well-recognized complication of peritonitis, especially in peritoneal dialysis patients. Of those patients that develop peritonitis, up to 15% have fungal peritonitis and 75% of these involve Candida spp.[2, 4] These patients often must permanently switch to hemodialysis, resulting in increased days of hospitalization and death rates.[2, 4, 8, 9] In one large multicenter study of fungal peritonitis secondary to peritoneal dialysis, only 22% of the fungal peritonitis episodes had concurrent bacterial sepsis, yet the poor outcomes of these patients were not significantly different between those with only fungal isolates and those with both bacterial and fungal isolates from their peritoneal fluid.
Tissue invasion by Candida is not seen in healthy animals. Consequently, histopathologic evidence of tissue infiltration indicates a transition of the organism from a commensal to an invasive pathogen. In our study, tissue invasion with fungal organisms was documented in case 4. Surgically obtained tissues were available for histologic examination in only 3 of our cases and a full necropsy was not allowed in any of the dogs that died. Therefore, it is possible that invasion may have been documented in other cases as well if more tissues were available for a histologic examination.
Some studies in people have shown contamination from the upper GI and biliary tract rather than the lower GI tract to represent an independent risk factor for Candida peritonitis.[1, 19] This seems to be true for dogs as well as all 5 cases in our study were associated with peritoneal contamination from the upper GI or biliary tract. In addition, the 2 previously reported cases in dogs were both contaminated from the small intestines.[5, 6]
Three dogs in our study were diagnosed with C albicans alone, one dog was infected with only C glabrata, and one dog had both C albicans and C glabrata. The 2 previously reported cases in dogs had cultures positive for C albicans only;[5, 6] thus, these are the first reported cases of C glabrata peritonitis in dogs. In people, Candida infection is the most common cause of fungal peritonitis and C albicans has traditionally been the predominant species isolated. However, C glabrata has recently increased in frequency and is considered the second most common cause of candidiasis.[8, 14, 20-23] Other commonly reported species include C parapsilosis, C tropicalis, and C krusei. In people, C glabrata is considered difficult to treat and is often resistant to many azoles, including fluconazole,[8, 12, 14, 17] which is usually the first line of treatment for invasive fungal infections. Our case 5, in which C glabrata was isolated alone, was treated with intravenous fluconazole and responded to treatment. In fact, both cases treated with intravenous fluconazole immediately after noting yeast organisms in cytologic specimens (cases 2 and 5) were the 2 dogs that survived. The potential clinical significance of this species of Candida isolated in veterinary patients will require further study.
Culture of Candida from peritoneal fluid is considered the gold standard for diagnosing candidiasis in peritonitis patients. However, culture requires at least 48 hours to provide a species identification.[12, 24] In people, direct cytologic examination of peritoneal fluid may provide an early diagnosis in almost 30% of cases. Beyond allowing an earlier diagnosis, this may also be of prognostic significance. One study in people showed identification of Candida on direct examination of peritoneal fluid to be an independent predictive factor for mortality. This likely reflects a greater concentration of organisms in the fluid and probably relates to true infection and proliferation of the organisms within the peritoneum. The Candida organisms in our cases were initially identified by cytologic examination of peritoneal fluid. Candida infection was not considered prior to this finding and fungal cultures were performed only after the microscopic identification of typical organisms in the peritoneal fluid. Thus, an index of suspicion in dogs with appropriate predisposing factors is relevant in making the diagnosis, and it is important for clinical pathologists to be aware of this potential.
Being cognizant of Candida's morphologic appearance may aid in early detection and may allow a tentative diagnosis of the organism via cytology. Candida spp. are typically seen in their yeast or blastoconidia form. With aqueous Romanowsky stains these are 3–8 μm round to oval, deeply basophilic organisms with a thin clear halo. Narrow-based budding may be seen. Their morphology is distinct from that of some other commonly identified pathogenic yeast, such as Histoplasma spp. or Blastomyces spp. All Candida spp. form blastoconidia or yeast, and most are dimorphic, forming germ tubes and then pseudohyphae. A few species of Candida, of which C albicans is the most common, are pleomorphic and have 3 vegetative growth forms: yeast, pseudohyphae, and true hyphae.[7, 25] Germ tube formation is the elongation of the yeast form and represents a transition to the hyphal form. Pseudohyphae are a succession of conjoined yeast, which display apparent constrictions at septal sites, while true hyphae do not have constrictions at septal sites and usually have even parallel sides. Pseudohyphae and true hyphae are noted usually in more established candidiasis and are not observed under normal physiologic conditions.[11, 25]
During an infection, Candida can readily react to the host's immune system by undergoing phenotypic switching, which changes its morphotype from blastoconidia to a more resistant germ tube or elongated yeast, pseudohyphae, and/or hyphal forms (Figure 1). Also, macrophage ingestion of C albicans can stimulate germ tube formation/elongation and allow escape and evasion of the immune system.[11, 16] This ability to alternate between yeast and hyphal morphotypes increases Candida's virulence and enhances its potential for tissue invasion.[11, 16] An exception is C glabrata, which appears to only exhibit the yeast pattern in vivo.
These findings suggest that Candida infection is a potential and deleterious complicating factor in dogs with peritonitis, particularly cases in which the infection is of enteric or biliary origin. The scarcity of prior reports suggests that this condition may be currently under-diagnosed or under-represented. This could result from either the failure to examine peritoneal fluid cytologically or inability to recognize the yeast in the sample as a true pathogen rather than a contaminant. It could also represent a possible increase in frequency of the infection, as noted in human studies, due to unnecessary broad coverage antimicrobial therapy, inappropriate or inadequate treatment, or prolonged therapy combined with the aforementioned predisposing factors. Of the 5 cases in this report, fungal organisms were identified upon cytologic examination of peritoneal effusion in all cases prior to a positive fungal culture. Two of the cases survived after early detection of yeast and immediate initiation of intravenous antifungal treatment. Identification of organisms in fluid samples can expedite appropriate treatment and potentially improve clinical outcome.
The authors thank Dr. Theresa Rizzi (Department of Pathobiology, Oklahoma State University) for her contribution in identifying the organisms. The authors also thank Dan Bradway, Washington Animal Disease Diagnostic Lab, for PCR analysis of the fungal cultures.