Sherpa.leeds.ac.uk

J Clin Pathol 2001;54:748–751
Evidence for antibiotic induced Clostridiumperfringens diarrhoea Abstract
Experimental diarrhoea has been produced in Clostridium diYcile is a well documented
human volunteers after oral administration of cause of antibiotic associated diarrhoea in
hospitalised patients, but may account for
implicated in the pathogenesis of antibiotic only approximately 20% of all cases. This
associated diarrhoea and sporadic diarrhoea.
leader reviews the current knowledge and
The aim of this article is to summarise the understanding of the pathogenesis, epide-
available evidence for C perfringens as a cause of miology, and diagnosis of non-food borne
antibiotic associated diarrhoea and provide an Clostridium perfringens diarrhoea. Al-
overview of the laboratory methods for diagno- though enterotoxigenic C perfringens has
sis of C perfringens diarrhoeal diseases.
been implicated in some C diYcile nega-
tive cases of antibiotic associated diar-
rhoea, C perfringens enterotoxin detec-
Evidence linking C perfringens to
tion methods are not part of the routine
non-food borne diarrhoeal disease
laboratory investigation of such cases.
In 1984, Borriello et al demonstrated that some Testing for C perfringens enterotoxin in
C diYcile toxin negative cases of antibiotic faecal samples from patients with anti-
associated diarrhoea were associated with biotic associated diarrhoea and sporadic
enterotoxigenic strains of C perfringens.2 In diarrhoea on a routine basis would have
their study, faecal specimens from patients with considerable
resource
implications.
antibiotic associated diarrhoea were examined.
Therefore, criteria for initiating investiga-
tions and optimum laboratory tests need
culture assay for CPEnt, and the presence of to be established. In addition, establishing
enterotoxin was further confirmed by an in the true burden of C perfringens anti-
biotic associated diarrhoea is important
(ELISA). All the cases of diarrhoea had high before optimum control and treatment
faecal counts (> 106 cfu/g) of C perfringens. measures can be defined.
Clostridium perfringens enterotoxin was detected (J Clin Pathol 2001;54:748–751)
in the stools of 11 patients with diarrhoea. Theseverity, duration, and sporadic nature of the Keywords: Clostridium perfringens; Clostridium diYcile; disease was not characteristic of C perfringens food poisoning, and most (n = 10) of thepatients developed diarrhoea after antibiotic Clostridium diYcile is the most commonly iden- tified pathogen in hospital acquired infective penicillins, cephalosporins, trimethoprim, and diarrhoea. However, for most cases (in some cotrimoxazole. Furthermore, CPEnt was not series up to 80%) the organism responsible is detected in the stools of patients with inflam- matory bowel disease (n = 29) and infective Clostridium perfringens are the most frequently diarrhoea (n = 12), or in patients with normal cited alternative causes of antibiotic associated stools (n = 16). These findings led the authors diarrhoea.1 Clostridium perfringens type A is an to conclude that enterotoxigenic C perfringens important cause of bacterial food poisoning was a possible cause of antibiotic associated world wide. There is mounting evidence to diarrhoea. Symptoms associated with non-food borne C perfringens diarrhoeal disease tend to strains can play a role in the aetiology of be protracted and more severe than those of diarrhoeal disease distinct from food poison- food poisoning.11 Common symptoms include ing, including antibiotic associated diarrhoea abdominal pain and diarrhoea, which are often and sporadic diarrhoea in humans.2–9 Clostrid- accompanied by blood and mucous in the fae- ium perfringens forms part of the normal human ces. Borriello and colleagues subsequently Department of
gut flora in small numbers—up to 103 colony confirmed their earlier findings by extending Microbiology, The
General Infirmary and
University of Leeds,
numbers of vegetative cells (> 108 enterotoxin Samuel et al studied the incidence of Old Medical School,
producing organisms) of C perfringens leads to diarrhoea associated with CPEnt by screening Leeds, LS1 3EX, UK
diarrhoeal illness, as in cases of food borne dis- diarrhoeal samples received from both general ease. The organisms multiply in the small practice and hospital patients.4 In addition to intestine and sporulate, releasing C perfringens screening for common enteric pathogens, sam- enterotoxin (CPEnt), which is responsible for ples were also tested for the presence of C perf- the classic symptoms of food poisoning. Faecal counts of > 106 C perfringens spores are seen in symptomatic cases, although similar counts shown to contain CPEnt accompanied by high can be obtained in patients without symptoms.
clostridial counts (> 103 cfu/g). The control Antibiotic induced Clostridium perfringens diarrhoea group samples, consisting of 120 normal stool 65 of the 212 faecal specimens examined. Most specimens, were negative for CPEnt and had of the enterotoxin positive cases were not asso- low faecal counts of C perfringens (< 103 cfu/g).
ciated with antibiotic treatment. In some of the The study demonstrated a clear association of cases, a single point source of food borne C perfringens diarrhoea with elderly hospital- organisms could not be excluded. In addition, ised patients, with 22 of 25 of the cases being a higher proportion (79%) of the enterotoxin hospital inpatients. Most (n = 18) of the positive group than the enterotoxin negative patients had received antibiotics before the onset of diarrhoea. The antibiotics associated importance of these findings remains uncer- with the cases were penicillins, cephalosporins, tain, but the surprisingly high prevalence of augmentin, erythromycin, trimethoprim, and enterotoxin positive cases raises questions nitrofurantoin. The clinical symptoms devel- about the specificity of the detection methods oped on diVerent days over a period of a few used. In a recent large study of infectious intes- weeks and were not typical of food poisoning. A tinal disease in the community in England, more recent study was designed to investigate CPEnt was found at a much lower prevalence the incidence of C perfringens as a cause of than in the study by Mpamugo et al, but was antibiotic associated diarrhoea.5 Hospitalised more frequently identified in general prac- patients were assigned to one of the four groups titioner patients (4%) than in the population on the basis of antibiotic usage and diarrhoeal Thus, evidence supporting a causal role for caused by other enteric pathogens were ex- C perfringens in non-food borne diarrhoeal dis- ease includes the presence of enterotoxin in the (n = 181) comprised patients who had or had faeces of patients with antibiotic associated not received antibiotics but had no diarrhoea.
diarrhoea or sporadic diarrhoea, as opposed to There was no predominant age group or sex.
asymptomatic individuals or individuals suVer- The stool samples were examined for the pres- ing from infectious intestinal disease caused by ence of C diYcile and C perfringens toxins.
other pathogens.13 14 It might be argued that CPEnt was detected in the faeces of 15 patients patients identified as suVering from C perfrin- with antibiotic associated diarrhoea, but in gens antibiotic associated diarrhoea were in fact none of the control groups. A further four cases of food poisoning. Although it is diYcult patients with diarrhoea had detectable CPEnt, to eliminate the possibility of food poisoning although there was no association with anti- completely, the clinical symptoms described biotic treatment. The possibility of cross infec- above were not consistent with those of food tion was suggested on the basis of geographical poisoning. There were no case clusters in rela- clustering, but isolates were not available for tion to time to suggest a common source food strain typing. However, an epidemiological borne outbreak. In a recent study, Collie et al investigation carried out elsewhere during an used restriction fragment length polymorphism outbreak of C perfringens diarrhoea showed that (RFLP) and PFGE to genotype 43 enterotoxi- infecting serotypes were present in the hospital genic C perfringens isolates.14 Results of their study showed that the enterotoxin gene (cpe) is patients.6 Clostridium perfringens serotypes re- extrachromosomal in C perfringens isolates as- contrast to the chromosomal enterotoxin gene samples. Only 9% of samples from areas not present in food poisoning isolates.14 The associated with symptomatic patients were importance of this observation is unclear, but it similarly positive. In a study from Japan, C per- does strengthen the possibility that C perfrin- fringens isolates recovered from the faeces of gens antibiotic associated diarrhoea is a distinct elderly hospitalised patients with sporadic entity. Furthermore, these findings may be useful for identifying reservoirs of genotypi- pulsed field gel electrophoresis (PFGE).7 Of cally distinct subpopulations of enterotoxigenic the 60 C perfringens isolates examined, 38 C perfringens isolates, which might help in instituting appropriate preventive measures.
miological and experimental data suggestedthat the diarrhoea was not related to a food Pathogenesis
borne outbreak, but was caused by a nosoco- There is a paucity of data relating to the patho- genesis of C perfringens antibiotic associated Sporadic cases of diarrhoea in the commu- diarrhoea. Use of antibiotics in elderly hospi- nity have also been caused by C perfringens, talised patients is considered the major risk particularly in the elderly population.8 9 In a factor for acquisition of the disease. It is not study by Mpamugo et al, faecal specimens from known whether antibiotic exposure primarily isolated sporadic cases of diarrhoea were permits the proliferation of small numbers of examined using a reverse passive latex aggluti- resident C perfringens strains or allows acquisi- nation (RPLA) kit (see below) for enterotoxin tion of enterotoxigenic C perfringens strains— detection, and positive results were confirmed for example, as a result of impaired colonisa- tion resistance. The symptoms are mediated by detected by ELISA in 43 of 45 specimens that the production of CPEnt, which is a 35 kDa yielded (in the RPLA kit) a diVerence of two polypeptide.14 15 The cpe gene has been cloned, wells or more between sensitised and control sequenced, and expressed in Escherischia coli.16 latex, and in 22 of 56 specimens with a one well Further studies are required to assess the diVerence. Hence, enterotoxin was detected in extent of genomic diversity among the food borne and non-food borne diarrhoeal disease group samples (n = 120) in their study, includ- isolates. These studies may also help to explain ing those from healthy elderly patients, had the diVerences in virulence between the geno- faecal counts of C perfringens > 103 cfu/g.
typically distinct subpopulations of enterotoxi- These conflicting results may be the result of genic C perfringens isolates. It is widely believed the low numbers of patients recruited in the that sporulation is essential for CPEnt produc- tion. However, Czeczulin et al reported that leagues21 and Stringer et al.22 Therefore, it is sporulation is not essential for CPEnt produc- essential to detect enterotoxin in faecal samples tion, although it does lead to an increased yield when attempting to confirm C perfringens as the of CPEnt.16 At present, regulation of CPEnt cause of illness.23 A large number of methods expression at the molecular level is not well including ELISA, RPLA, and tissue culture understood. Studies using gene probe assays assays have been used for the detection of have shown that cpe is present in only 6% of C perfringens enterotoxin in faeces, but each has global C perfringens isolates, whereas 60% of limitations.24 25 Tissue culture assay, using vero the isolates associated with food poisoning cells, is least sensitive (40 ng enterotoxin/g fae- were found to be cpe positive by hybridisa- ces) and reproducible among toxin detection tion.17 18 These findings were surprising consid- methods.24 An in house ELISA available at the ering the role of CPEnt in the pathogenesis of Public Health Laboratory Service central food C perfringens food poisoning. Further studies hygiene laboratory has been suggested as the identified sequence diVerences in cpe regions gold standard among toxin detection methods that had been used to design some of the DNA on the basis of its sensitivity (5 ng enterotoxin/ g), specificity, and reproducibility.13 24 The interfered with probe hybridisation, resulting in RPLA kit (Oxoid, Basingstoke, UK) has been false negative conclusions about cpe positive shown to be both sensitive and reproducible, isolates. It is also possible that some of the iso- but non-specific interference by faecal matter lates tested may have been non-enterotoxigenic has been reported. In addition, some authors normal flora C perfringens isolates because only have reported that this method has relatively single colonies were selected for testing.19 More poor sensitivity (50–100 ng enterotoxin/g).25 recently, use of gene probe assays based on the cpe sequencing work of Czeczulin and col- ELISA kit (produced by TechLab; distributed leagues16 found that only a small proportion of C perfringens isolates carried the cpe gene.20 Of commercially available. However, data com- 454 fresh C perfringens isolates from a variety of paring this ELISA with the methods referred to sources (animal and environmental), only 16 above are not available. In one study,5 detection (3.5%) were polymerase chain reaction (PCR) of enterotoxin alone using the new ELISA kit was used as the laboratory diagnostic criterionfor inpatients with antibiotic associated diar- Diagnosis
rhoea. A study undertaken at our hospital using The optimum laboratory investigations for the commercial ELISA kit found that 15% of diagnosis of C perfringens antibiotic associated faecal specimens from patients with antibiotic diarrhoea have not been established, and the associated diarrhoea were CPEnt positive (NJ absence of a “gold standard” test hinders Ransome et al, Abstracts of 38th Interscience further progress. The tests that have been used by previous investigators involve the direct Chemotherapy, 1998, San Diego. Abstract K detection of faecal enterotoxin, quantitative 126). This figure is very similar to the results of stool cultures for C perfringens, and absence of a previous study.5 Ten percent of isolates from other enteric pathogens in stool specimens.
Culture of C perfringens alone does not prove positive by PCR (NJ Ransome et al, Abstracts the diagnosis, particularly in light of the high numbers of asymptomatic carriers in the insti- tutional setting.21 22 However, stool cultures San Diego. Abstract K 126). It is possible that provide the isolates from which useful data can cpe positive C perfringens isolates produced enterotoxin in vivo to cause disease, but there Clostridium perfringens can be found in small numbers (up to 103 cfu/g) in the faeces of CPEnt to be detected by ELISA, thus generat- healthy individuals.4 However, there have been ing false negative results. This could be reports of high faecal counts of C perfringens compounded by the collection of samples late (> 107 cfu/g) in specimens obtained from after the onset of illness, because there is healthy geriatric patients.21 22 In a study carried evidence that maximum excretion of entero- out by Yamagishi et al,21 persistently high num- toxin occurs early in the food borne diarrhoeal bers of C perfringens were found in the faeces of Japanese geriatric patients (five of 30). Stringer confirm that these findings hold true in cases of et al studied faecal carriage of C perfringens in non-food borne diarrhoeal illness, and to younger and elderly long stay hospital patients determine the importance of these preliminary and found similar results in five of 21 elderly findings. In addition 13 of 19 presumptive patients (counts > 107 cfu/g).22 However, the C perfringens isolates from ELISA positive fae- younger long stay patients in the same hospital cal specimens were cpe negative (NJ Ransome had faecal counts in the range of 103 to 104 cfu/ et al, Abstracts of 38th Interscience Conference g. These findings are in contrast to the findings on Antimicrobial Agents and Chemotherapy, reported by Samuel et al.4 None of the control 1998, San Diego. Abstract K 126). It is possible Antibiotic induced Clostridium perfringens diarrhoea that these specimens contained very low num- considerable resource implications. In addi- bers of cpe positive organisms and conse- tion, establishing the true burden of C perfrin- quently tested isolates were PCR negative.
gens antibiotic associated diarrhoea is impor- Furthermore, C perfringens isolates can lose the tant before optimum control and treatment determine the enterotoxigenicity of C perfrin- 1 Wilcox MH. Infective antibiotic associated diarrhoea. In: gens isolates and thereby avoid the problems Wilcox MH, ed. Infection highlights 1999. Oxford: HealthPress, 2000:57–64.
associated with culture, and subsequent en- 2 Borriello SP, Larson HE, Welch AR, et al. Enterotoxigenic terotoxin production in vitro, and the inherent Clostridium perfringens: a possible cause of antibiotic-associated diarrhoea. Lancet 1984;i:305–7.
limitations of toxin detection methods. In an 3 Larson HE, Borriello SP. Infectious diarrhoea due to attempt to overcome these potential problems, Clostridium perfringens. J Infect Dis 1988;157:390–1.
4 Samuel SC, Hancock P, Leigh DA. An investigation into direct detection of cpe in primary faecal speci- Clostridium perfringens enterotoxin associated diarrhoea.
mens by PCR was undertaken in our hospital, J Hosp Infect 1991;18:219–30.
5 Hancock P. Antibiotic associated diarrhoea C. diYcile or C.
using a PCR protocol based on that of Kokai- perfringens? Reviews in Medical Microbiology 1997;8(suppl
Kun et al.20 However, all tested faecal samples, 6 Borriello SP, Barclay FE, Welch AR, et al. Epidemiology of including the specimens spiked with entero- diarrhoea caused by enterotoxigenic Clostridium perfrin- toxigenic C perfringens reference strains, pro- gens. J Med Microbiol 1985;20:363–72.
7 Wada A, Masuda Y, Fukayama, et al. Nosocomial diarrhoea duced negative results, despite universal bacte- in the elderly due to enterotoxigenic Clostridium perfrin- gens. Microbiol Immunol 1996;40:767–71.
8 Brett MM, Rodhouse JC, Donovan TJ, et al. Detection of amplification reactions. It is assumed that the Clostridium perfringens and its enterotoxin in cases of spo- high concentration of background bacterial radic diarrhoea. J Clin Pathol 1992;45:609–11.
9 Mpamugo O, Donovan T, Brett MM. Enterotoxigenic DNA in stool samples may have interfered with Clostridium perfringens as a cause of sporadic cases of amplification. However, elsewhere PCR assays diarrhoea. J Med Microbiol 1995;43:442–5.
10 Skjelkvale R, Uemura T. Experimental diarrhoea in human have been used successfully for the direct volunteers following oral administration of Clostridium detection of enterotoxin genes in faeces.27 28 For perfringens enterotoxin. J Appl Bacteriol 1977;43:281–6.
11 Borriello SP. Clostridial diseases of the gut. Clin Infect Dis example, Saito et al successfully used the 1995;20:S242–50.
supernatant of sporulation cultures of faeces as 12 Tompkins DS, Hudson MJ, Smith HR, et al. A study of infectious intestinal disease in England: microbiological template DNA to detect cpe by PCR.28 Further findings in cases and controls. Commun Dis Public Health work is needed to overcome the problems asso- 1999;2:103–13.
13 Bartholomew BA, Stringer MF, Watson GN, et al. Develop- ciated with the extraction and amplification of ment and application of an enzyme linked immunosorbent C perfringens DNA from stool samples. Resolu- assay for Clostridium perfringens type A enterotoxin. J ClinPathol 1985; 38:222–8.
14 Collie RE, McClane BA. Evidence that the enterotoxin gene important tool for understanding more about can be episomal in Clostridium perfringens isolates associ-ated with non-food borne human gastrointestinal diseases.
C perfringens diarrhoeal diseases.
J Clin Microbiol 1998;36:30–6.
15 McClane BA, Strouse RJ. Rapid detection of Clostridium perfringens type A enterotoxin by enzyme linked immuno-
sorbent assay. J Clin Microbiol 1984;19:112–15.
16 Czeczulin JR, Hanna PC, McClane BA. Cloning, nucleotide sequencing, and expression of the Clostridium perfringens Conclusion
enterotoxin gene in Escherichia coli. Infect Immun 1993;61:
It has been 17 years since Borriello et al first demonstrated the role of enterotoxigenic C per- 17 Van Damme-Jongsten M, Wernars MK, Notermans S. Clon- ing and sequencing of the Clostridium perfringens entero- fringens in the aetiology of antibiotic associated toxin gene. Antonie van Leeuwenhoek 1989;56:181–90.
diarrhoea or sporadic diarrhoea, but only 18 Van Damme-Jongsten M, Rodhouse MJ, Gilbert RJ, et al.
Synthetic DNA probes for detection of enterotoxigenic limited further research has followed. Current Clostridium perfringens strains isolated from outbreaks of knowledge and understanding of the pathogen- food poisoning. J Clin Microbiol 1990;28:131–3.
19 Rood JI, McClane BA, Songer JG, et al, eds. The Clostridia: esis, epidemiology, and diagnosis of C perfrin- molecular biology and pathogenesis. Cambridge: Cambridge gens non-food borne diarrhoeal illness is based 20 Kokai-Kun JF, Glenn Songer J, Czeczulin JR, et al.
on the results of a few published studies, and Comparison of western immunoblots and gene detection partly on the extrapolation of data obtained assays for identification of potentially enterotoxigenicisolates of Clostridium perfringens. J Clin Microbiol from C perfringens food poisoning outbreaks.
1994;32:2533–9.
Clostridium diYcile is a well documented cause 21 Yamagishi T, Serikawa T, Morita R, et al. Persistent high numbers of Clostridium perfringens in the intestines of of antibiotic associated diarrhoea in hospital- Japanese aged adults. Jpn J Microbiol 1976;20:397–403.
22 Stringer MF, Watson GN, Gilbert RJ. Faecal carriage of Clostridium perfringens. J Hyg (Lond) 1985;95:277–88.
approximately 20% of all cases.29 Although 23 Birkhead G, Vogt RL, Heun EM, et al. Characterisation of enterotoxigenic C perfringens has been impli- an outbreak of Clostridium perfringens food poisoning byquantitative faecal culture and faecal enterotoxin measure- cated in some of these C diYcile negative cases ment. J Clin Microbiol 1988;26:471–4.
of antibiotic associated diarrhoea, CPEnt 24 Berry PR, Rodhouse JC, Hughes S, et al. Evaluation of ELISA, RPLA and vero cell assay for detecting Clostrid- detection methods are not part of the routine ium perfringens enterotoxin in faecal specimens. J Clin laboratory investigation of such cases. The cri- Pathol 1988;41:458–61.
25 Brett MM. Kits for the detection of some bacterial food poi- teria for initiating investigations and optimum soning toxins: problems, pitfalls and benefits. J Appl Micro- laboratory tests need to be established. In biol 1998;84(suppl):110S–18S.
26 Brynestad S, Synstad B, Granum PE. The Clostridium per- addition, any diagnostic method used should fringens enterotoxin gene is on a transposable element in be rapid, easy to perform, and inexpensive. The type A human food poisoning strains. J Clin Microbiol
1997;143:2109–15.
27 Paton AW, Paton JC, Goldwater PN, et al. Direct detection needs to be further evaluated against other of Escherichia coli shiga like toxin genes in primary faecalcultures by polymerase chain reaction. J Clin Microbiol methods before being adopted by routine diag- 1993;31:3063–7.
nostic laboratories. Testing for C perfringens 28 Saito M, Matsumoto M, Funabashi M. Detection of Clostridium perfringens enterotoxin gene by the polymer- enterotoxin in faecal samples from patients ase chain reaction amplification procedure. Int J Food with antibiotic associated diarrhoea and spo- Microbiol 1992;17:47–55.
29 Wilcox MH. Treatment of Clostridium diYcile infection. J radic diarrhoea on a routine basis would have Antimicrob Chemother 1998;41(suppl C):41–6.

Source: http://sherpa.leeds.ac.uk/archive/00000256/01/wilcoxmh3.pdf