Incidence and characterization of Clostridium perfringens isolated from antibiotic-associated diarrhoeal patients: a prospective study in an Indian hospital

https://doi.org/10.1016/j.jhin.2005.12.015Get rights and content

Summary

Clostridium perfringens has been reported as causing between 2–15% of all cases of antibiotic-associated diarrhoea (AAD), and may be diagnosed by detection of enterotoxin in faeces. A prospective study comprising 150 diarrhoeal patients and 100 non-diarrhoeal controls was undertaken to assess the incidence of C. perfringens-associated diarrhoea in an Indian hospital. Methods used included C. perfringens culture, reverse passive latex agglutination (RPLA) and enzyme-linked immunosorbent assay (ELISA) for detection of enterotoxin, and polymerase chain reaction (PCR) assay for the presence of enterotoxin gene. Attempts were made to type the isolates by multiplex PCR. Of the 150 diarrhoeal stool samples tested, 13 were culture positive. Of these, four were positive for C. perfringens enterotoxin by RPLA, two were positive by PCR and two were positive by RPLA and ELISA. Twenty-seven samples were positive for culture of C. perfringens in non-diarrhoeal controls but none were positive for enterotoxin either by RPLA or by PCR. The average incidence of C. perfringens AAD using these methods was 2.6%. Toxin typing showed that all the isolates belonged to type A. To conclude, the relatively low incidence of toxigenic C. perfringens suggests that enterotoxigenic C. perfringens is not a major cause of AAD in this population.

Introduction

Antibiotic-associated diarrhoea (AAD) is a serious complication of treatment for infection, and the pathogenic role of toxin-producing strains of Clostridium difficile is well established.1, 2 However, for most cases (up to 80% in some series), the organism responsible is undiagnosed. Although many other different aetiological agents have been proposed, it remains unclear which species are the most significant causes of disease. Clostridium perfringens and Staphylococcus aureus are the most frequently cited alternative causes of AAD.3

C. perfringens was first implicated as a cause of AAD in 19844 when free faecal enterotoxin was detected in 11 patients with diarrhoea following antibiotic administration. The evidence for a causal role of C. perfringens in AAD has increased and it has been reported to account for between 2% and 15% of all cases.5, 6 C. perfringens enterotoxin is a 35-kDa protein produced constitutively in vivo at low levels by bacteria during the vegetative growth phase, but expression is upregulated during sporulation.7, 8 C. perfringens enterotoxin binds to the gut epithelial cells where it is internalized and alters cell membrane permeability, with resultant epithelial cell death leading to diarrhoea. Unlike C. difficile infection, C. perfringens AAD does not result in the formation of pseudomembranes.4 C. perfringens enterotoxin is only found in faeces of symptomatic or infected people and is absent from faeces of convalescent or healthy subjects.9 For determining whether a foodborne outbreak is due to C. perfringens, several epidemiological studies have suggested the detection of enterotoxin in faecal specimens as a marker of C. perfringens diarrhoea. Recent surveys have suggested that only a low percentage (∼5%) of all C. perfringens isolates carry the cpe gene that encodes enterotoxin.10, 11

C. perfringens has been classified into five toxigenic types (A–E) on the basis of its ability to produce exotoxins: four major lethal toxins (alpha, beta, epsilon and iota) other than minor toxins and enterotoxin.12, 13 The alpha toxin is believed to be the major factor responsible for pathological changes of gas gangrene in humans and necrotic enteritis and enterotoxaemia in animals.14 The beta toxin is mainly associated with necrotic enteritis in humans and cattle.15, 16 The epsilon toxin is mainly responsible for acute enterotoxaemia in animals, and the iota toxin is mainly responsible for enterotoxaemia in calves and sheep.17 C. perfringens strains of biotypes A, B, C, D and E are producers of alpha toxin; alpha, beta and epsilon toxins; alpha and beta toxins; alpha and epsilon toxins; and alpha and iota toxins, respectively.13 Of these, types A and C are commonly involved in human infections whereas types B, D and E are of major importance in veterinary medicine. Biotype determination has a particular value for detection of pathogenic C. perfringens strains and for epidemiological studies. Determination of the presence and type of toxin produced by each C. perfringens strain is based on in vivo methods with the use of laboratory animals, in vitro techniques on synthetic media, immunoenzymatic methods, and molecular diagnostic techniques.10, 18, 19, 20 However, they demand large amounts of active toxin and a large number of laboratory animals.13 Recently, molecular methods based on hybridization and/or amplification were introduced for characterization of C. perfringens strains.20, 21, 22, 23, 24

To date, no information exists on the prevalence of C. perfringens-associated diarrhoea in India, although it is well established in the West. This prompted the authors to perform a prospective study to assess the incidence of C. perfringens-associated diarrhoea in their hospital and to type the isolates genotypically by multiplex polymerase chain reaction (PCR).

Section snippets

Specimens

A total of 250 specimens were analysed for C. perfringens. This represented 150 specimens from patients with AAD submitted for C. difficile testing at the Anaerobic Laboratory, Microbiology Department, All India Institute of Medical Sciences, and 100 non-diarrhoeal specimens as controls. The patients were evaluated prospectively by grouping them into the following three categories: hospitalized patients with diarrhoea (N = 150); hospitalized patients without diarrhoea (N = 58); and non-hospitalized

Results

Demographic data on age, sex and duration of diarrhoea for the study group are summarized in Table I. In total, 1.6% (4/250) of stool specimens were positive for C. perfringens enterotoxin by RPLA. Of the four toxin-positive samples, two were negative for cpe gene.

Discussion

In the authors' hospital, enterotoxigenic C. perfringens was responsible for 1.3% (2/150) of cases of diarrhoea. These cases came from oncology medical wards. This compares with the 8% prevalence rate of C. difficile infection in the same group. Evidence for antibiotic-induced C. perfringens diarrhoea has also been shown in several studies.4, 6, 27 The prevalence of C. perfringens enterotoxin in stool samples in these studies was reported to vary between 1.6% and 6.4%. In the present study,

Acknowledgement

The authors thank Dr E. Augustynowicz, Department of Sera and Vaccine Evaluation, National Institute of Hygiene, Warsaw, Poland for providing the reference strains of C. perfringens. This work was supported by the Indian Council of Medical Research, New Delhi, India.

References (30)

  • J.R. Czeczulin et al.

    Regulated expression of Clostridium perfringens enterotoxin in naturally cpe-negative type A, B, and C isolates of C. perfringens

    Infect Immun

    (1996)
  • American Public Health Association

    Clostridium perfringens food poisoning

  • M. Van Damme-Jongsten et al.

    Cloning and sequencing of the Clostridium perfringens enterotoxin gene

    Antonie Van Leeuwenhoek

    (1989)
  • J.F. Kokai-Kun et al.

    Comparison of Western immunoblots and gene detection assays for identification of potentially enterotoxigenic isolates of Clostridium perfringens

    J Clin Microbiol

    (1994)
  • E.P. Cato et al.

    Genus Clostridium

  • Cited by (0)

    View full text