Transmission of tuberculosis by patients with HIV infection

BMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6967.1515 (Published 03 December 1994) Cite this as: BMJ 1994;309:1515
  1. G M Scott,
  2. J Holton
  1. Infection control doctor, Infection control doctor, Clinical Microbiology, University College London Hospitals, London WC1E 6AU.

    EDITOR,—R J Kent and colleagues used restriction fragment length polymorphism analysis to show that three strains of Mycobacterium tuberculosis from patients positive for HIV at this hospital were indistinguishable.1 The results became available to us only one year later. Our investigation could not identify a plausible route of transmission.

    Case 1—The index patient (a 25 year old South American man) was on a ward remote from the AIDS unit between 15 and 25 June. He had fever, a dry cough, weight loss, and shadowing of the right middle lobe but with neutropenia and lymphopenia and antibodies to HIV. He was transferred on 25 June to a side room on AIDS ward A and walked through AIDS ward B to have a bronchoscopy on 26 June, returning on a trolley. Acid fast bacilli were seen in the bronchoalveolar lavage fluid, so treatment was started and he was isolated as a potential source of infection for at least two weeks; he was discharged on 12 August.

    Case 2—The second patient (a 29 year old African woman), who had cytomegalovirus retinitis, was on AIDS ward B (separated from ward A by a corridor) from 24 June to 21 July and from 23 August to 11 October. Cultures of induced sputum (obtained on 26 August), bronchoalveolar lavage fluid (2 September), blood (6 September), and pleural fluid (10 September) yielded M tuberculosis.

    Case 3—The third patient (a 54 year old English man), who also had cytomegalovirus retinitis, was on AIDS ward B transiently on 3, 26, and 29 June and then from 18 to 29 July and from 14 to 19 August, when he died. M tuberculosis was isolated from blood and faeces obtained on 17 August.

    If this was an outbreak these events imply that the patients in cases 2 and 3 caught tuberculosis from the patient in case 1. So far as can be established, however, apart from walking through ward B on 26 June (which is unlikely to have led to transmission2), the patient in case 1 was isolated on a different ward. The patients were unaware of each other and could not have caught tuberculosis from, say, a common staff source because the patient in case 1 was admitted with active tuberculosis and AIDS was diagnosed later. Staff are not shared between the two AIDS wards. Bronchoalveolar lavage was performed with different instruments many weeks apart by different endoscopists. The patients in cases 2 and 3 did not enter the endoscopy room on 26 June, and this room has negative pressure ventilation.

    The epidemiology does not support the molecular biology findings and leaves us with worrying uncertainty about how best to manage patients with tuberculosis on AIDS wards. More rapid diagnosis would help, but patients infected with HIV may contract infection from people living in the same room in whom sputum smears yield negative results but sputum cultures yield positive results.3 Currently we isolate patients with changes evident on chest radiography and those in whom sputum smears or sputum cultures yield positive results. None of our side rooms have negative pressure ventilation. The suggestion of unexpectedly high transmissibility of tuberculosis in patients with AIDS suggests that even more strict protocols should be adopted, including the isolation of all patients with undiagnosed respiratory disease in AIDS in negatively ventilated side rooms. The emergence of drug resistant tuberculosis would make this even more important.


    1. 1.
    2. 2.
    3. 3.
    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
    Sign up for a free trial