Transmission of tuberculosis in British centre for patients infected with HIVBMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6955.639 (Published 10 September 1994) Cite this as: BMJ 1994;309:639
- R J Kent,
- A H C Uttley,
- N G Stoker,
- R Miller,
- A L Pozniak
- Regional Tuberculosis Centre, Dulwich Public Health Laboratory, London SE22 8QF
- Department of Clinical Sciences, London School of Hygiene and Tropical Medicine, London WC1E 7HT Middlesex Hospital, London W1N 8AA
- Department of Genitourinary Medicine and Department of Medicine, King's College School of Medicine and Dentistry, London SE5 9PT
- Correspondence to: Dr Pozniak.
- Accepted 27 April 1994
Patients infected with HIV are at particular risk of tuberculosis, because of both the reactivation of latent infection and the rapid development of disease after exposure to any new source of infection. Many outbreaks of tuberculosis in patients infected with HIV and in the carers of those patients have occurred in hospitals and other institutions in the United States and Europe.1, 2 Those with tuberculosis caused by multiple drug resistant strains have been associated with a high mortality.
Study of the epidemiology of tuberculosis has been helped by restriction fragment length polymorphism analysis with the insertion sequence IS6110. This repetitive element is randomly inserted, usually in multiple copies, in the genomic DNA of members of the Mycobacterium tuberculosis complex. A standardised methodology for this technique has been proposed.3 We used this methodology to investigate a cluster of cases of tuberculosis in a British care centre for patients infected with HIV.
Method and results
We examined 96 strains of Mycobacterium tuberculosis that had been isolated from 69 patients infected with HIV and had been referred to the regional tuberculosis centre at Dulwich Public Health Laboratory between August 1991 and June 1993. We compared the strains with control strains from 84 patients not infected with HIV that had been referred by the same laboratories during the same period. The source laboratories comprised all but one of those that served the care centres for patients infected with HIV in south east England at the time of the study.
DNA was extracted from a fresh subculture of each strain and digested with restriction endonuclease Pvu II. DNA fragments were separated by agarose gel electrophoresis and blotted on to nylon membrane. After hybridisation with an IS6110 probe labelled with digoxigenin, fragments containing this sequence were detected by chemiluminescence. The patterns obtained were compared visually.
All the strains contained at least one IS6110 sequence. Except for three pairs and one cluster of four cases, all from the patients infected with HIV, the strains isolated from different patients had different patterns of restriction fragment length polymorphism. We reviewed the case notes of the four patients who had strains with no distinguishable differences (table).
The first three cases are epidemiologically related: two patients (case 2 and 3) were in a ward with a patient (case 1) whose sputum smear test yielded a positive result and in whom bronchoscopy was performed in the ward. The two patients both had low CD4 lymphocyte counts and developed disseminated tuberculosis within several weeks of exposure to the other patient. The relationship between these three patients and a fourth (case 4) is not known. The latter was a man who had never been married, had no known contact with hospital 1, and resisted attempts at contact tracing. An intermediate patient whose strain was not referred to us was perhaps involved.
This small cluster of cases shows the rapid progression and dissemination of mycobacterial infection in patients with advanced HIV disease. The outbreak would not have been recognised without restriction fragment length polymorphism analysis. Recognition and investigation of future outbreaks might be helped if all strains of tuberculosis are routinely stored and typed at reference laboratories.
Tuberculosis that is acquired in a hospital might be prevented by attention to infection control precautions designed to limit the spread of airborne pathogens. Guidelines have been published by the Centers for Disease Control, Atlanta,4 and the British Thoracic Society.5 Probably the most important factors are the early recognition of tuberculosis and the adequate isolation of cases. Also, procedures likely to generate aerosols, such as bronchoscopy, should be performed in a suitable environment. Units that care for patients infected with HIV should review their procedures and policies for preventing and managing tuberculosis. Conversion of such units to the standards recommended mended by the Centers for Disease Control might be expensive but should be considered carefully.
The equipment used in this study was bought with a grant to ALP from King's College School of Medicine and Dentistry. NGS was funded partly by the Overseas Development Administration.