An algorithm for tuberculosis screening and diagnosis in people with HIV

Area of Science:

• Infectious Disease Epidemiology
• Public Health
• Clinical Diagnostics

Background:

• Tuberculosis (TB) screening is crucial for individuals with human immunodeficiency virus (HIV) to enable early diagnosis and safe initiation of antiretroviral therapy (ART) and isoniazid preventive therapy (IPT).
• Current guidelines for TB screening in HIV-infected populations are lacking, despite the common practice of screening for chronic cough.

Purpose of the Study:

• To develop and validate an effective algorithm for screening and diagnosing tuberculosis in patients with HIV infection.
• To compare the characteristics of patients diagnosed with TB against those without TB to inform screening strategies.

Main Methods:

• Consecutive enrollment of HIV-infected patients from outpatient clinics in Cambodia, Thailand, and Vietnam.
• Collection of multiple biological samples (sputum, urine, stool, blood, lymph-node aspirate) for mycobacterial culture.
• Comparative analysis of patient characteristics to derive a diagnostic algorithm.

Main Results:

• Tuberculosis was diagnosed in 15% of 1748 patients.
• Screening for cough (2-3 weeks) had low sensitivity (22-33%), while a combination of cough (any duration), fever (any duration), or night sweats (≥3 weeks) showed 93% sensitivity and 36% specificity.
• A negative diagnosis was aided by sputum smears, normal chest radiograph, and CD4+ count ≥350 cells/mm³; definitive diagnosis often required mycobacterial culture.

Conclusions:

• Tuberculosis screening in HIV-infected individuals should incorporate a combination of symptoms, moving beyond solely assessing chronic cough.
• Negative screening results for a combination of key symptoms suggest safe initiation of ART and IPT.
• Mycobacterial culture remains essential for diagnosing TB in most patients who screen positive or have indeterminate results.