Introduction

Since the outbreak of serious illness and deaths due to Clostridium novyi that occurred amongst injecting drug users (IDUs) in 2000 [1], the Anaerobe Reference Laboratory (ARL) at the University Hospital of Wales in Cardiff has been alerted to the role of clostridial infections in this group of patients. It is believed that, somewhere along the supply chain, heroin is being contaminated with clostridial spores. These spores have been shown to survive the heroin ‘cooking up’ process that commonly involves heating in citric acid (pH 2.1) prior to injection [2]. Only skin or muscle ‘poppers’, those IDUs who inject subcutaneously or into muscle tissue rather than a vein seem to be affected, as the injectate creates a localised necrotic focus that is suitable for the germination of the clostridial spores. Mixed clostridial spores may be present including C. tetani and some cases of clinical tetanus in IDUs have been reported, but an unusual number of infections primarily due to a rarely isolated clostridial species has recently come to our attention.
Methods and Results
The first such case was in December 2003 when a 35 year old female IDU from Glasgow presented at hospital with a necrotic lesion at an injection site in her buttock. The organism isolated from this lesion was referred to the Anaerobe Reference Laboratory for identification and was identified according to the phenotypic criteria of Holdeman et al [3] as Clostridium histolyticum (FIGURE 1). In the next few weeks, further indistinguishable isolates from IDUs from other cities across England led to an alert being issued by the Health Protection Agency Communicable Disease Surveillance Centre in the Communicable Disease Report Weekly [4]. Later, over a four month period to April 2004, another eleven isolates referred to the ARL from injection site infections or blood cultures taken from heroin injecting drug users across England and Scotland have been identified as C. histolyticum. Molecular typing methods were applied to these isolates to determine their relatedness using unrelated strains from the National Collection of Type Cultures (NCTC) culture collection and one wild clinical isolate from a crushed-hand injury for comparison. To the best of our knowledge, no molecular typing methods have ever been applied to this organism as it is so rarely isolated from clinical material.
Figures 2 and 3 show the results of pulsed-field gel electrophoresis (PFGE) and polymerase chain reaction (PCR) ribotyping analysis respectively, of DNA extracted from the twelve IDU isolates and comparator strains. All the IDU isolates of C. histolyticum were indistinguishable by both methods and showed different DNA profiles to the NCTC strains. The one wild isolate detailed above also had a distinct profile.

Discussion

Referrals of C. histolyticum from human material are so rare that previously the ARL had received just one clinical isolate of C. histolyticum in the preceding 20 years from an infected crushed-hand injury in an agricultural worker.
C. histolyticum is a member of the gas-gangrene group of clostridia that may be isolated from soil, bone-meal and gelatin. It produces potent exotoxins that have proteolytic and necrotising properties causing severe localised necrosis. However, these toxins do not elicit the systemic effects that caused such dramatic loss of life as seen in the C. novyi - associated outbreak in 2000.
The referrals had originated from hospitals in nine different towns or cities around England and Scotland including Glasgow, London, Brighton, Manchester, Middlesbrough, Banbury, Liverpool, Derby and Nottingham. The patient demography was 2:1, females to males, and the average age was 35 years. This ratio is at odds with the usual demography of IDUs in which males usually predominate and infection may be related to a higher ratio of skin and muscle ‘poppers’ among women who have difficulty injecting into a vein. The patients had mostly presented at Accident and Emergency Units where debrided material was sent for microbiological investigation to the on-site diagnostic bacteriology departments. Ten referrals were isolates from such material and two were from blood cultures. In some cases, mixed clostridia were isolated and the local laboratories then referred these unusual isolates to the ARL for identification as is common practice.
Although the isolates are from as far apart as London and Glasgow, the results of this typing investigation suggests a common source to these infections. The most probable scenario is a batch of heroin that was contaminated quite early in the production or supply chain prior to distribution within the United Kingdom (UK). Interestingly, this outbreak appears only to be affecting the UK as, to date, there have been no reports of C. histolyticum infections in IDUs in other countries. The outbreak appears to be ongoing as we have received several more referrals since April 2004 and this intelligence needs to be cascaded to drug support workers in the field and also to medical staff, particularly in accident and emergency units. Diagnostic microbiology departments should also be alerted to infections in IDU’s presenting with severe local necrosis at injection sites and pay attention to any unusual clostridia isolated. UK drug support organisations highlighted this problem in 2000 during the C. novyi outbreak advising users not to inject heroin into tissues if at all possible and this warning should be repeated. Other European countries should also be alert in case supply routes of contaminated heroin alter.

Acknowledgements

We thank the departments of clinical diagnostic microbiology in the relevant hospitals for referral of the isolates identified as Clostridium histolyticum.