Early Empirical Tuberculosis Treatment in HIV-Positive Patients Admitted to Hospital in South Africa: An Observational Cohort Study

Abstract

Background Empirical TB treatment in HIV-positive inpatients is common and may undermine the impact of new diagnostics. We sought to describe empirical TB treatment and compare characteristics and outcomes with patients treated for TB after screening. Methods Retrospective observational cohort study of HIV-positive inpatients treated empirically for TB prior to TB screening. Data on clinical characteristics, investigations and outcomes were collected from medical records. Comparison cohorts with microbiologically-confirmed or empirical TB treatment after TB screening with Xpert MTB/RIF and urine lipoarabinomannan assays were taken from South African STAMP trial site. In-hospital mortality was compared using a competing-risks analysis adjusted for age, sex and CD4 count. Results Between January 2016 and September 2017, 100 patients excluded from STAMP were treated for TB empirically prior to TB screening. After enrolment in STAMP and TB screening, 240/1177 (20.4%) patients received TB treatment, of whom 123 had positive TB tests and 117 were treated empirically. Characteristics were similar among early empirically treated patients and those treated after TB screening. 50% of early empirical TB treatment was based on radiological investigations, 22% on cerebrospinal or pleural fluid testing, and 28% on clinical features alone. Only 11/100 empirically treated patients had subsequent microbiological confirmation. In-hospital mortality was lower in patients with microbiologically-confirmed TB compared to those treated empirically (adjusted sub-distribution HR 0.5, 95% CI 0.3-0.9). Conclusions Empirical TB treatment remains common in severely ill HIV-positive inpatients. These patients may benefit from TB screening using existing rapid diagnostics, both to improve confirmation of TB disease and reduce over-treatment for TB.

Design and protocol for a cluster randomised trial of enhanced diagnostics for tuberculosis screening among people living with HIV in hospital in Malawi (CASTLE study)

“…TB is extremely difficult to fully refute, as evidenced by both autopsy findings (half of all TB diagnosed among PLHIV in autopsy series meta-analysis was not diagnosed antemortem despite the people who died having been in contact with health services prior to death) [ 4 , 14 , 15 ], and that in clinical practice TB is often treated “empirically” (i.e. without microbiological confirmation) [ 17 ]. Anticipating that this intervention may increase the amount of empiric TB treatment used, we chose to have one study sputum culture as a minimum objective TB reference, whilst acknowledging that a single sputum culture result is far from a definitive way to diagnose or refute TB.…”

Addressing TB-related mortality in adults living with HIV: a review of the challenges and potential solutions

“… 101 However, an observational cohort study of TB/HIV co-infected patients conducted in South Africa found that in-hospital mortality was lower in patients with microbiologically confirmed TB, compared with patients treated empirically (HR: 0.5; 95% CI: 0.3–0.9), emphasizing the importance of trying to obtain a microbiological diagnosis of TB in PLHIV. 46 …”

“…45 However, since the risk of smear-and Xpert-negative TB is also increased in advanced HIV, culture plays an important role in confirming TB in PLHIV, and it has been reported that mortality is higher in patients without culture-confirmed TB. 45,46 Diagnosis -biomarker-based tests. LAM.…”

“…101 However, an observational cohort study of TB/HIV co-infected patients conducted in South Africa found that in-hospital mortality was lower in patients with microbiologically confirmed TB, compared with patients treated empirically (HR: 0.5; 95% CI: 0.3-0.9), emphasizing the importance of trying to obtain a microbiological diagnosis of TB in PLHIV. 46 Optimal treatment for TB meningitis. There is increasing evidence that we are underdosing RIF in many people with TB and that higher RIF doses are safe and well tolerated.…”