Equine Colitis X Associated with Infection by Clostridium Difficile NAP1/027

Abstract

A 14-year-old Quarter Horse with a 48-hr history of colic was euthanized after failure to respond to treatment. At necropsy, cecal and colonic mucosae were congested throughout, and there was segmental edema and significant thickening of the intestinal wall. Excessive numbers of mononuclear cells were found in mucosal lamina propria. Submucosal hemorrhage was diffuse and extensive, and Clostridium difficile toxins A and B were detected. Large numbers of C. difficile were isolated, and genetic characterization revealed them to be North American pulsed-field gel electrophoresis type 1, polymerase chain reaction ribotype 027, and toxinotype III. Genes for the binary toxin were present, and toxin negative–regulator tcdC contained an 18-bp deletion. This genotype comprises the current human “epidemic strain,” which is associated with human C. difficile–associated disease of greater than historical severity. The diagnosis was peracute typhlocolitis, with lesions and history typical of those attributed to colitis X.

Colitis X, so-called because of its lack of a defined etiology, has been described at intervals over the past 50 years. 12,26 It has been associated with antimicrobial treatment 12,23,30 and with high protein–low cellulose diets, though many cases have had no apparent predisposing factors. Salmonella spp. and Escherichia coli have been suggested as causative agents, 17 but most of the attention has focused upon clostridia, namely Clostridium perfringens and Clostridium cadaveris. It has now become apparent that many C. difficile–affected adult horses present with classic clinical signs, and particularly the gross and microscopic lesions, of colitis X. A typical fatal case of colitis X in an adult horse with the current human epidemic strain of C. difficile (ribotype 027/toxinotype III) as the likely etiologic agent is described.

A 14-year-old Quarter Horse with a 48-hr history of colic and diarrhea was euthanized after treatment with ceftiofur (for putative salmonellosis) and worsening of clinical signs. At necropsy, the animal was in good overall condition. A large amount of yellow to clear fluid was found in the abdomen, and the lungs were somewhat congested. The small intestine contained excessive fluid contents, but no mucosal lesions were observed. The cecum and colon contained normal semisolid ingesta, but the colonic and cecal mucosae were very congested throughout. Segmental edema was observed, as was significant thickening of the intestinal wall. The mucosae were intact in most areas, but 2-to 4-mm ulcers were found in the distal portion of the colon.

Intestinal tissues fixed in 10% phosphate buffered formalin were embedded in paraffin, sectioned, and stained with hematoxylin and eosin and Gram stains. Microscopic examination revealed that the colonic submucosa was edematous, resulting in separation of tissues. Infiltrations of primarily mononuclear inflammatory cells throughout the submucosa and extending into mucosa resulted in excessive numbers of mononuclear cells in mucosal lamina propria. Vascular congestion was common throughout. In some areas, focal hemorrhages were found in the submucosa just beneath the muscularis mucosa, while in other areas submucosal hemorrhage was diffuse and extensive. Numerous large Gram-positive rods were observed on the mucosa and extending into the lamina propria.

Colonic and cecal contents were examined for toxins A and B by use of a commercial enzyme immunoassay. a Toxins were present in substantial amounts, exceeding 4+ on the scale of the assay. Affected tissues were examined by bacteriologic culture. Samples of intestinal content and mucosal scrapings were plated on taurocholate cefoxitin cycloserine fructose agar 38 and streaked for isolation. Incubation was in an environment of 5% H², 5% CO², and 90% N² for 48 hr at 37°C. Clostridium difficile was identified by colony morphology, the “horsebarn-like” odor, yellow-green fluorescence under long-wave ultraviolet illumination, a negative indole reaction, and production of L-proline aminopeptidase. b Identity was confirmed by polymerase chain reaction (PCR) amplification of genes for toxins A (tcdA) and B(tcdB), using primers reported elsewhere. 9,33 The colon and cecum yielded large numbers of C. difficile. The diagnosis was peracute enteritis involving the colon and cecum, with lesions and history typical of colitis X.

Isolates were examined by PCR ribotyping, 25 pulsed-field gel electrophoresis (PFGE), 16 toxinotyping, 25 and PCR assays for binary toxin gene cdtB 32 and deletions in tcdC. 19 Findings of these assays identified the isolates as belonging to North American PFGE type 1 (NAP1), PCR ribotype 027, and toxinotype III. Gene cdtB was present, and tcdC contained an 18-bp deletion. These characteristics are consistent with those of the current epidemic strain, a hypervirulent strain affecting humans worldwide. 8

Colitis X is an acute disease of horses that occurs sporadically and is characterized by abdominal pain, severe watery to hemorrhagic diarrhea, dehydration and electrolyte imbalances, and toxic shock. 12,17,23,31 It is usually diagnosed postmortem, where cecal and colonic mucosal edema and hemorrhagic necrosis, as well as watery to bloody intestinal contents, are found. 31 Disseminated intravascular coagulation, colonic mucosal necrosis, and anarchic multiplication of bacteria in devitalized intestinal segments are seen microscopically and are considered pathognomonic. 12,26,35 Differential diagnoses include salmonellosis, heavy metal intoxication, and occlusive verminous arteritis. 12

A causative organism for colitis X has not been definitively identified, and, in fact, the etiology is likely multifactorial. Factors affecting disease severity include age, immunocompetence, and the degree to which normal intestinal defenses and microflora are disrupted. 34 High protein–low cellulose diets and stresses such as transport, concurrent illness, or surgery may play a role. However, the risk of developing colitis X is most significantly associated with antimicrobial therapy. Horses pretreated orally with clindamycin and lincomycin and inoculated with intestinal content from horses with fatal idiopathic colitis developed colitis X. 22 Tetracyclines, lincomycin, and erythromycin have been implicated in predisposition of horses to colitis X. 1,10,23,30 Milder forms of antibiotic-associated colitis can also occur. 10

The search for a microbial etiologic component has focused on members of the genus Clostridium. Horses developing colitis X in response to inoculation with antimicrobials and intestinal contents from horses with natural disease yielded an organism resembling C. cadaveris. 22 Production of disease by inoculation with pure cultures of this organism has not been reported, and its possible role remains equivocal. Clostridium perfringens has been suggested to be an etiologic agent, but evidence is lacking. Intravenous injection of C. perfringens enterotoxin (CPE) produced colic, severe shock, and pathological changes typical of colitis X, 21 but enterotoxigenic strains are seldom recovered from cases (unpublished data). Furthermore, CPE is produced in the gut, and there is no evidence that it enters circulation. Large numbers of C. perfringens have been found in intestinal contents of horses with fatal colitis after administration of oxytetracycline and erythromycin, 1,10 but there is no evidence to support an etiologic connection. Thus, the role of these species remains controversial. 34

Recent findings strongly suggest that C. difficile may be the most important, if not the sole, player in etiology of colitis X. Clostridium difficile is a confirmed pathogen of many mammals. 27 Human disease has been known for more than 30 years 4 and varies in severity with the infecting strain. 20 Since around the year 2000, the disease has emerged in piglets 28,29 and calves. 11,24 Disease manifests as typhlocolitis in most species, and both incidence and severity vary by host, age, and antibiotic exposure. Differences in colonization rates and toxin-receptor densities have been proposed as explanations for variable susceptibility of species and age groups within a species. 15

Clostridium difficile causes hemorrhagic, necrotizing enterocolitis with high mortality in foals. 2 The organism and its toxins are very common in feces of diarrheic foals but uncommon in normal foals. 13 Disease occurs in the absence of antimicrobial therapy, but treatment with erythromycin or gentamicin in combination with rifampicin also led to excretion of C. difficile. 7

Clostridium difficile–associated diarrhea of variable severity occurs sporadically and epidemically in adult horses. It is often antibiotic associated or nosocomial, 6,7,18 and the role of the organism in its etiology has been confirmed via a prospective study. Clostridium difficile was isolated from >12% of horses with colitis but from only 0.4% of normal adults. Toxins were detected in 21.8% of diarrheic adults but in only 1.2% of normal adults. Among patients with colitis, toxin-positive horses are less likely to survive than toxin-negative horses. 37

Clostridium difficile or its toxins were demonstrated in feces of 40% of mature horses that developed acute colitis after treatment with beta-lactam antibiotics. 5 Acute colitis caused by C. difficile has occurred in the dams of foals treated with erythromycin and rifampicin for Rhodococcus equi pneumonia. 7

A wide variety of genotypes of C. difficile have been isolated from horses with fatal colitis. 3,14 The striking feature about the isolate described in the current case is that it is NAP1, PCR ribotype 027, toxinotype III, which is the single most important epidemic strain causing human C. difficile–associated disease in North America, the United Kingdom, and Europe. 20,36

This finding is of unknown significance for human or even equine health. The ultimate source of ribotype 027 epidemic strains is unknown, although they have been found in pigs in Canada (J.S. Weese, personal communication, 2007) and calves in the United States (M.P. Hailey, M.A. Anderson, and J.G. Songer, unpublished data, 2008). It is possible that the human epidemic strain arose in domestic animals, but there is no evidence to support this position. Given that animals typically maintain a rather relaxed and philosophical attitude toward sanitation, they may, by fecal to oral cycling, increase the population of C. difficile spores to which humans can be exposed. This deserves further study. It is also possible that ribotype 027 strains of C. difficile are of greater virulence for horses than historical strains and that this may lead to more severe disease. This should perhaps be investigated by a more detailed investigation of cases, including characterization of isolates.