Elsevier

The Lancet

Volume 354, Issue 9185, 2 October 1999, Pages 1159-1163
The Lancet

Articles
Unrecognised Mycobacterium tuberculosis bacteraemia among hospital inpatients in less developed countries

https://doi.org/10.1016/S0140-6736(98)12325-5Get rights and content

Summary

Background

Nosocomial transmission of Mycobacterium tuberculosis is a global public-health concern. Although early clinical recognition of M tuberculosis in hospital inpatients is critical for effective infection control, such recognition may be difficult in patients with HIV infection. To find out whether M tuberculosis bacteraemia frequently goes unrecognised, we did a prospective blood-culture survey in an infectious-diseases hospital in Thailand and a general hospital in Malawi.

Methods

Consecutive febrile (⩾37·5°C axillary or ⩾38·0°C orally) hospital inpatients (aged ⩾18 years) were enrolled; blood was obtained for mycobacterial culture and HIV testing. Simple diagnostic tests, such as chest radiographs and sputum smears, were ordered by clinicians as deemed necessary, and were carried out with existing local resources.

Findings

Of 344 patients enrolled, 255 (74%) were HIV infected, the median age was 33 years (range 18–87), and 208 (61%) were male. 34 (10%) patients had Mtuberculosis bacteraemia; five of these patients were already on antituberculosis therapy. Only HIV-infected patients had M tuberculosis bacteraemia. Of the 29 patients with M tuberculosis bacteraemia who were not already receiving antituberculosis therapy, 13 (45%) had an abnormal chest radiograph or a positive sputum smear. 16 (55%) patients had no additional diagnostic test results to indicate M tuberculosis infection; 18 (81%) of these had a cough.

Interpretation

In less developed countries where both M tuberculosis and HIV infections are prevalent, M tuberculosis bacteraemia may frequently go unrecognised among febrile hospital inpatients.

Introduction

Mycobacterium tuberculosis is a growing, worldwide public-health threat. About a third of the world's population is infected; M tuberculosis causes up to three million deaths annually.1 The burden of active tuberculosis is spread disproportionately between two key regions: southeast Asia and sub-Saharan Africa. Of about 40 million episodes of active tuberculosis that will have occurred in young adults worldwide between 1990 and 2000, more than 28 million will have occurred in southeast Asia, and 9 million in sub-Saharan Africa.1

The HIV pandemic has contributed to the increased spread of tuberculosis; HIV-infected patients are 30 times more likely to develop active tuberculosis and become infectious than are HIV-negative individuals.1 Of the estimated 20 million individuals with HIV infection or AIDS in the world in 1997, more than five million were diagnosed in southeast Asia, and 14 million in sub-Saharan Africa.2 In these regions, where both HIV and M tuberculosis infections are prevalent, the hospital setting may have an important role in the spread of M tuberculosis.

Prevention of nosocomial transmission of M tuberculosis requires early clinical recognition of patients with active disease so that isolation precautions can be taken and therapy promptly initiated.3 However, recognition of tuberculosis can be difficult in HIV-infected patients; many such patients present with atypical signs and symptoms, and they are susceptible to various other pulmonary infections that can mimic tuberculosis and make diagnosis difficult.4, 5, 6, 7, 8 Although simple diagnostic tests, such as chest radiographs9 and sputum smears,7 may assist in the recognition of active disease in hospital inpatients, resources for carrying out such diagnostic tests are limited in many less developed countries.

Mycobacteraemia is a key event in the pathogenesis of tuberculosis, and is detected in many HIV-infected patients with active tuberculosis.10, 11, 12M tuberculosis is now being recognised as an important cause of bacteraemia in less developed countries.13, 14 However, clinicians in these regions frequently do not suspect underlying mycobacteraemia, and therefore do not request appropriate blood cultures. In addition, lack or inappropriate use of clinical microbiology resources in hospitals in less developed countries, or the absence of clinical algorithms for making the decision to obtain M tuberculosis blood cultures from patients suspected of having mycobacteraemia, may contribute to low detection rates of this disease.

We did a blood-culture survey in hospitals in southeast Asia and sub-Saharan Africa to measure the prevalence of M tuberculosis bacteraemia among febrile, adult inpatients; to find out whether active tuberculosis in febrile patients with M tuberculosis bacteraemia is, in practice, frequently unrecognised; and to assess factors associated with the under-recognition of active disease in patients with M tuberculosis bacteraemia.

Section snippets

Study population

The surveys were carried out in an infectious-disease hospital near Bangkok, Thailand, during February and March, 1997, and at a government-run, general hospital in Lilongwe, Malawi, during August and September, 1997. Approval of the study protocol was obtained from the institutional review boards at the Centers for Disease Control and Prevention (GA, USA) and at the hospitals in Thailand and Malawi.

We recruited consecutive adult patients (aged ⩾18 years) who were febrile (oral temperature

Results

344 patients (216 in Thailand and 128 in Malawi) were enrolled at the two sites (table 1). Of those originally approached, no patients in Thailand, and four in Malawi declined to participate. A greater percentage of male patients was enrolled in Thailand (68%) than in Malawi (48%; p<0·001). 255 (74%) patients were HIV infected. 34 (10%) patients had M tuberculosis bacteraemia.

Clinical predictors of M tuberculosis bacteraemia in the combined study population included the presence of HIV

Discussion

Much current emphasis on controlling the global increase in tuberculosis is placed on promoting directly observed treatment, short course (DOTS). The main objective of this prevention effort, led by the WHO,1 is to achieve increased cure rates through greater compliance with antimicrobial therapy. This emphasis is appropriate given that both HIV-infected and non-infected patients may be rendered non-infectious for tuberculosis through treatment, and that the emergence of multidrug resistance in

References (22)

  • RW Shafer et al.

    Extrapulmonary tuberculosis in patients with human immunodeficiency virus infection

    Medicine

    (1991)
  • Cited by (81)

    • Opsonic monoclonal antibodies enhance phagocytic killing activity and clearance of Mycobacterium tuberculosis from blood in a quantitative qPCR mouse model

      2019, Heliyon
      Citation Excerpt :

      In sub-Saharan Africa, many HIV positive individuals are co-infected with TB [2, 3]. In HIV-positive patients, TB is the leading cause of death [2], and many develop MTB bacteremia, sepsis and disseminated disease [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15], often dying within 30 days of hospitalization [10, 15]. Furthermore, strains of MTB are becoming increasingly resistant to antibiotic therapy, and the emergence of multiple drug resistant (MDR) and extensively drug resistant (XDR) strains have increased the urgency to find new approaches for prevention and treatment of TB [16].

    View all citing articles on Scopus
    View full text