Baseline Xpert MTB/RIF ct values predict sputum conversion during the intensive phase of anti-TB treatment in HIV infected patients in Kampala, Uganda: a retrospective study

Juliet Namugenyi,Joan Kalyango,Emmanuel Sendaula,Jan Fehr,Joseph Musaazi,Yukari C Manabe,Andrew Kambugu,Willy Ssengooba,Achilles Katamba,Christine Sekaggya-Wiltshire,Barbara Castelnouvo

doi:10.1186/s12879-021-06220-6

Abstract

BackgroundIn resource-limited settings, sputum smear conversion is used to document treatment response. Many People living with HIV (PLHIV) are smear-negative at baseline. The Xpert MTB/RIF test can indirectly measure bacterial load through cycle threshold (ct) values. This study aimed to determine if baseline Xpert MTB/RIF could predict time to culture negativity in PLHIV with newly diagnosed TB.MethodsA subset of 138 PLHIV from the ‘SOUTH’ study on outcomes related to TB and antiretroviral drug concentrations were included. Bacterial load was estimated by Mycobacterium Growth Indicator Tubes (MGIT) culture time-to-positivity (TTP) and Lowenstein Jensen (LJ) colony counts. Changes in TTP and colony counts were analyzed with Poisson Generalised Estimating Equations (GEE) and multilevel ordered logistic regression models, respectively, while time to culture negativity analysed with Cox proportional hazard models. ROC curves were used to explore the accuracy of the ct value in predicting culture negativity.ResultsA total of 81 patients (58.7%) were males, median age 34 (IQR 29 ̶ 40) years, median CD4 cell count of 180 (IQR 68 ̶ 345) cells/μL and 77.5% were ART naive. The median baseline ct value was 25.1 (IQR 21.0 ̶ 30.1). A unit Increase in the ct value was associated with a 5% (IRR = 1.05 95% CI 1.04 ̶ 1.06) and 3% (IRR = 1.03 95% CI 1.03 ̶ 1.04) increase in TTP at week 2 and 4 respectively. With LJ culture, a patient’s colony grade was reduced by 0.86 times (0R = 0.86 95% CI 0.74 ̶ 0.97) at week 2 and 0.84 times (OR = 0.84 95% CI 0.79 ̶ 0.95 P = 0.002) at week 4 for every unit increase in the baseline ct value. There was a 3% higher likelihood of earlier conversion to negativity for every unit increase in the ct value. A ct cut point ≥28 best predicted culture negativity at week 4 with a sensitivity of 91. 7% & specificity 53.7% while a cut point ≥23 best predicted culture negativity at week 8.ConclusionBaseline Xpert MTB/RIF ct values predict sputum conversion in PLHIV on anti-TB treatment. Surrogate biomarkers for sputum conversion in PLHIV are still a research priority.

Highlights

In resource-limited settings, sputum smear conversion is used to document treatment response
This study sought to determine if baseline ct values predict time to sputum culture conversion and correlate with decreasing bacterial load measured by Mycobacteria Growth Indicator Tube (MGIT) time to positivity (TTP) and Löwenstein–Jensen (LJ) colony counts during the first two months of treatment in Human Immunodeficiency Virus (HIV) infected participants with severe immunosuppression
We found that the baseline Xpert Mycobacterium tuberculosis (MTB)/RIF ct value was significantly associated with; 1) change in bacterial load measured by TTP and colony counts up to 4 weeks following treatment initiation, 2) eventual time to culture negativity in the first two months of treatment

Summary

Introduction

In resource-limited settings, sputum smear conversion is used to document treatment response. The Xpert MTB/RIF test (Cepheid, Sunnyvale, CA) is widely used in TB diagnosis [9] due to its improved sensitivity over smear microscopy and a short turnaround time [6, 7, 10, 11] It provides a measure of bacterial load through cycle threshold (ct) values [12]. Shenai et al in South Africa [5] demonstrated that baseline ct values provided a good estimate of time to sputum culture conversion and likelihood of relapse in an HIV-negative population. This study sought to determine if baseline ct values predict time to sputum culture conversion and correlate with decreasing bacterial load measured by Mycobacteria Growth Indicator Tube (MGIT) time to positivity (TTP) and Löwenstein–Jensen (LJ) colony counts during the first two months of treatment in HIV infected participants with severe immunosuppression. We explored the sensitivities and specificities of various ct cut offs to confirm how correct the predictions of negativity are at different time points of treatment

Objectives

Methods

Results

Conclusion