Intended for healthcare professionals


HIV associated tuberculosis

BMJ 1997; 314 doi: (Published 28 June 1997) Cite this as: BMJ 1997;314:1847

A barometer for wider tuberculosis control and prevention

  1. Richard Coker, Consultant physiciana,
  2. Rob Miller, Senior lecturerb
  1. a Departments of Genitourinary Medicine, HIV, and Respiratory Medicine, St Mary's Hospital NHS Trust, London W2 1NY
  2. b Division of Pathology and Infectious Diseases, University College London Medical School, London WC1E 6AU

    People who are infected with HIV are at an increased risk of contracting tuberculosis. The WHO estimates that just over 20 million people are currently infected with HIV and of these 6 million are co-infected with Mycobacterium tuberculosis. Worldwide there has been a resurgence of tuberculosis, mainly in developing countries but also in the United States and Europe. Between 1987 and 1993 tuberculosis rates increased by 35.5% in London (with the increase most notable in inner London) compared with 15% in England and Wales as a whole.1 2 However, it is unclear to what extent the prevalence of HIV associated tuberculosis has increased in the capital, largely because notification of tuberculosis in the HIV infected population is unreliable and probably underestimates the problem.3 Nosocomial spread has occurred in several specialist HIV centres in Britain and elsewhere, and drug (and multidrug) resistance is a growing problem.4 5 Because of the shortened incubation period for tuberculosis, HIV infection highlights weaknesses in tuberculosis control programmes.

    Recent data from London have shown that HIV infected patients originating from Africa are at increased risk of developing tuberculosis, yet they are rarely offered preventive treatment.6 Furthermore, these patients may be unaware of their HIV status at presentation.6 7 Notification rates in black Africans have more than doubled recently, from 60.4/100 000 in 1988 to 135.2/100 000 in 1993. In contrast, rates have fallen in West Indians, people from the Indian subcontinent, and white people (J Watson, personal communication). With the high prevalence of HIV infection in parts of Africa and the global spread of HIV to other areas endemic for tuberculosis, tuberculosis clinics should be encouraged to increase testing for HIV.8

    In New York City the prevalence of tuberculosis is four times the American national rate, with nosocomial spread and multidrug resistant tuberculosis (that is, tuberculosis resistant to isoniazid and rifampicin with or without resistance to other antimicrobials) reaching a peak in the early 1990s. Enhanced infection control programmes in institutions (along with directly observed therapy) have contributed substantially to the more recent rate reductions.9 Microbial typing by restriction fragment length polymorphism can improve the detection of nosocomial spread and help detect clusters in the community.10 However, there is now greater possibility of litigation if procedures for infection control are not followed. Recent British recommendations encouraging a coordinated local approach are welcome but will need significant additional resources if they are to be adopted.11

    Until recently, Britain has experienced only low levels of drug resistant tuberculosis. Between 1982 and 1991 only 0.6% of primary isolates were resistant to isoniazid and rifampicin.12 Provisional figures for 1994 in London, however, show 2.8% of isolates from north Thames and 1.4% of isolates from south Thames region are multidrug resistant.13 Recent evidence suggests the risk of drug resistance and, in particular, multidrug resistance is significantly greater in the HIV infected population, particularly among those who have been treated for tuberculosis before.4

    Directly observed therapy for tuberculosis reduces the prevalence of acquired drug resistance, relapse, and, perhaps in the long term, primary drug resistance.5 In Britain, as in the United States, it is recommended for “non-compliant patients, and those likely to be non-compliant.”14 However, compliance with antituberculous treatment is notoriously unpredictable. In Britain there is a national shortage of clinical nurse specialists in tuberculosis, and district nurses may be unwilling or unable to take on this service. As a result, directly observed therapy cannot be implemented effectively.

    HIV associated tuberculosis is a barometer for tuberculosis control. Efforts should be concentrated on early identification of patients with infectious tuberculosis, early initiation of effective treatment, and effective screening of patients with tuberculosis for HIV and vice versa. Improved control of tuberculosis in this group and in the population as a whole will depend on improved facilities for isolating patients with infectious tuberculosis and wider use of directly observed therapy. Additional resources and funding are needed.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    View Abstract