Yersinia infection with Clostridium difficile colitis

Antibimic associated diarrhea appears to be largely due to Closcridium difficile and may have, at least in part. a toxin mediated pathogenesis. Because a poor correlation exists between measurable toxin titres and symptoms, additional copathogenic factors may be important. All patients seen during a two year period with diarrhea within four weeks of antibiotic therapy, a positive $tool culture for C difficile and a positive stool cytotoxin assay specific for C difficile were investigated. All patients had stools cultured for cnteric bacterial pathogens including Salmonella , Shigella. Campylobacter, Aeromonas and Yersinia species. Seven patients had Yersinia species isolated during the course of their illness; no other entcric pathogens were identified . In four patients, Y enrerocolitica was cultured simultaneously with C difficile prior to treatment, and in one of these, Y fredriksenii was also isolated. Of six patients with persistent or recurrent sympcoms after treatment for C difficile ( ie, vancomycin in five and mecronidazole in one patient), four had positive yersinia cultures at the conclusion of therapy (Y enrerocolicica in three and Y fredriksenii in two patients). All but one of these patients had been yersinia culture negative prior to therapy for C difficile Patients with and without yersinia isolates were then compared with respect to age, sex, clinical symptoms and sigmoldoscopic as well as rectal biopsy findings. The presence of yersinia was associated with male sex, younger age and abdominal pain; other features including hematochezia, fever, arthralgia , cytotoxin titre , sigmoidoscopy and rectal biopsy could not distinguish patients with and without Yc.-rsinia species. Thus, ycrsinia may be associated with an antibiotic related diarrheal illness usually attributed to C difficile alone and may be observed in the setting of persistent or recurrent symptoms following treatment for C difficile diarrhea. Can J Gastro• enterol 1989;3( 1 )121-25

C OLITIS MAY COM PLICATE ANTIM lcrobial therapy and, in the vast majority of patie nts, appears to be due co Clos1ridium difficile ( l ). In most cases, treatm ent wi th oral vancomyci n is effective altho ugh o ne o r mo re symp tomatic rela pses may occur ( 1). Failure to e radicate the organism, possibly due to spo rularion, or acq uisi tion of a new strain of C difficile, may be respo nsible (2). Alternatively, coinfection with ano ther organism or acqu isition of a second o rganism could occu r.
In the present stud y, all patien ts seen in the hospi tal d uring a two year period with C diffici le associated colitis were evaluated prior to and at the conclusion of therapy fo r a second bacte rial pathogen or parasite. In thi s specific setting, a significan t propo rtion of patients wi th disease initially attributed to C diffici le were observed co have Yersinia species infection , bu t no o the r, bacte ria l pathogens.

MATERIALS AND METHODS
Specimen source: Fecal s pecime ns we re de ri ved fro m in pa ti e nts b e in g investigated fo r diarrhea or fro m patients referred to an ou tpatient clinic specializing in gastroi ntestina l diseases; a small numbe r of specimens were su bmitted from a un ive rsity student health service. During the two year period of this study, 40 3 stool specime ns fro m 269 patients were cultured for C difficile and its cytotox in; a ll stool speci me ns were ro uti nely culture d for Yersmia species. A total of 14 patients were identified with diarrhea beginning within fou r wee ks of antibiotic administration associated with a positive stool culture and cytotoxin assay for  provoquces par C difficile. La colite qui compliquc parfois les thera pies antimicrobiennes semble attribuable, clans la grande maiorite des cas, a C difficile Le plus sou vent, un traitemcnt a base de vancomycine par voie orale suffit, bien qu'une rcchute symptomatique ou plusieurs puissent se produire. Si !'on ne parvient pas a eliminer l'organisme, ii est possible qu'il y ait sporulation ou qu'une nouvelle souche de C difficile soit rcsponsable de l'echec; ou encore, qu'il y ait co-infection par un autre organisme ou acquisition d'un second agent. Dans l'etude presente, qui porte sur une pcriodc de dcux ans, rous les patients attcints d 'une colitc associee a C difficile ont ete evalues avant et au terme d 'une thcrapie destince a combattre un second agent pathogene ou parasite. Dans ce cadre particulier, un nombre significatif des patients atteints de maladies tout d'a bord attribuces a C difficile sc sont averes porteurs de Yersinia, mais d'aucun autre agent bacterien pathogene C diffici le. O f these, no patie nt developed diarrhea in hospital and all were referred fr o m within British Columbia. Srool specime ns we re obta ine d fro m all patients at the onset a nd conclusion of therapy fo r diarrhea presumed to be caused by C difficile infec tion. Speci mens were submitted in contai ne rs with Cary-Blair transport medium or fresh withou t transport med ium . co mpanied by addi tional fecal samples to exami ne fo r the presence of in testinal parasites. No fecal sa mples were examined for e nteric viruses. Yersinia culture and de tection me• thods: Fecal mate ria l was streaked on yersinia selective {Schiemann C IN) medium (Gibso, Madison . Wisconsin) and incubated at 35°C for 18 to 24 h afte r cold e nrichment in te rvals o f 24 h , three days and again after one and two weeks.
Yersinia species were fi rst detected by their characte ristic colonial morphology appearance on ye rsi nia selection C IN medium . As d escribed e lsew he re (3), typical colonies have a 'b ullseye' appea ra nce with a deep cherry red centre and transpare nt ma rgins. Colon ies also have a ground glass appearance under the stereo microscope and h ave a distin ctive odour. Suspect yersinia isolates were furthe r screened using a routine triple sugar iron agar slant fo r de tection of glucose, lactose a nd sucrose fe rm e n tation or hydrogen gas production, as well as an urea slant fo r urease prod uction. In addition , lysine an d indo lc test reactions as well as motil ity after incubation at 28°C were assessed . Fur ther confir mation of Yers mra se ries was done using an AP! 20E enterobacteriaccae system (API Laboratory Prod ucts Inc, St Laurent, Quebec), incu bated at 28°C. A typical group o f sugar fe rmentation reactions for yersin ia permitting initial distinction of the diffe rent species is tabulated in Table I. All positive isolates ide nti fied to species were subm itted to an independent reference laboratory for confirmation as well as definition of biotype and seroty pe of the organism Clostridium difficile: C difficile was cultured anaerobically on CD and Columbia CNA media for 48 h . {Capco Anaerobic Environmental Syste m, Santa Clara, California). Iden tification was based on colonial appcarnnce , typica l 'barnya rd odour', Gra m sta in morphology, aerobic and anaerobic subcultu re and anae robic A PI sugar ferme ntation profile. Cytotoxin assay followed the me thod of Allen (4) u sing filte red srool su per nata nt (1 0,000 g ; 0 .4 5 µm fil ter pore size), human foreskin fi b roblast cultures and C difficile antitoxin (TD Wilkins, Vi rginia Polytechn ic Institute, Blacksburg, Virginia). Titres we re d eterm ined using serial 10-fo ld dilutio ns of the su pernatant.

RESULTS
Clinical features: Table 2   rhea. Inclusion criteria required the presence of diarrhea within four weeks of antibi otic th erapy; one patient with bloody di arrhea h ad a negative yersinia cultu re , one had a positive initial culture and cine with initially negative cultures. whose diarrhea became bloody after failed treatment, subseq uently haJ yersin ia isolated . Abdominal pain or tenderness at the time of initial culture was more common when yersin ia was present ( three of four patients. 75",;,) than when yersi nia was absent (two of seven patients. 29'~[,). With fa iled therapy for C difficile, one patient had persistent abdominal pain and persistence of Y enteroc.:olitica in stool. In one patient with persiste nt d iarrhea and initially negative cultures for yersinia, abdominal pain and the presence ofboth Y enterocolitica anJ Y Jrcdriksenii developed after vancomycin th erapy for C difficile. Duration of symptoms did not differ in yersinia positive patients compared co yersinia negative patients.   Table 4 ). The endoscopic and histologic observations did not differentiate patients with ycr~inia 1wlated in stool cultures compared to patients wtthout positive cultu res . The patient with persisting Y entcrocolwca bdore and after treatment has a 'nonspecrfrc' cnlnis histologically confirmed at hoth times wtth positive cultures.

Clostridium difficile cytotoxin titres:
Tahle 5 shows C Jrj/ICtlc cytotoxin titres in patients wtth and without positive yersinra isolates. No differences m toxm titres were observed All four patients with pcrsbient diarrhea and positive ye rsinia cultures after therapy had C difficile cytotoxin titres measured; all of these patil'nts had negative cytotoxin assays after therapy. Two other patients with persistent symptoms fo llowing treatment had negauve ycrsmia cultures; one of these patients h ad a persistently positive C difficile stool cytotoxin assay ( titres of 1/ 10.000 before and after a course of vancomycin therapy).

DISCUSSION
Several organisms have been etiologically implicated in post antibiotic colitis. Staphylococcus aureu.,1 received much attention initially ( 5) and. more recently. C perfnngcns has bt·en suggested as a possible cause (6) However. the majority of cases are presen tl y ascnhed to C difficile C difficile produces an entemtoxin (ie, toxin A) and a potent cytomxin ( re, cox in  B) (7)(8)(9). The cytotoxin may induce typical changes in the colonic mucosa (8) and its detection has been used to establish the diagnosis (I). However, there appears m be a poor correlation between cyrotoxin titre (toxin B) and the severity of the clinical illness ( 10). Although toxin A may correlate more strongly with symptoms, other factors may also be involved ( 11) Of 108 patients with post antibiotic C diffiale colitis in Sweden, none had other enteric pathogens isolated; some patients, however, had detectable cymtoxin but negative cultures for C difficile ( 12). This raised the possibility of C difficile cytotoxin production by undetected organisms, perhaps other than C difficile ( 12, 13) Bartlett (14)   here. This seemed to be especially associated with persisting or recurrent diarrhea.
Clinical differentiation of patients with and without yersinia in the presence of C difficile appears to be difficult. Abdominal pain was more common with yersinia in the present study, but this symptom is well recognized in patients with C diffiale colitis ( 18). Fever anJ bloody diarrhea were also present in both groups. Arthralgias, reported in up to 30% of patients with Y enterocolirica infection ( 19,20) also occurs in C difficile colitis ( 21-24) and could not predict the presence of Yersinia species in the present patient:;.
Yersinia enterocolitis has been associated with normal mucosa, non:;pecific colitis and colonic mucosa! aphrhoid ulcerations (25), but in this study, sig-moidoscopic examination and rectal mucosa! biopsy of patients with C difficile diarrhea could not differentiate between patients with and without yersinia coinfection. Pseudomembranes were seen in a minority of patients wrth and without yersinia. Two of the present patients, both having isolates of C difficile and Y encerocolicica, presented as an exacerbation of previously diagnosed inflammatory bowel disease; both C difficile (26)(27)(28) and yersinia ( 3,29) have been implicated in this setting. Under these circumstances, morphologic diagnosis may be very difficult.
In summary, an inordinately high prev-