Abstract:Background: There are few prospective data about the incidence and mortality associated with pulmonary tuberculosis in intensive care units (ICUs), and none on the accuracy and clinical effect of the Xpert-MTB/RIF assay in this setting. We aimed to measure the frequency of culture-positive tuberculosis in ICUs in Cape Town, South Africa and to assess the performance and effect on patient outcomes of Xpert MTB/RIF versus smear microscopy for diagnosis of tuberculosis. Methods: We did a prospective burden of disease study with a randomised controlled substudy at the ICUs of four hospitals in Cape Town. Mechanically ventilated adults (≥18 years) with suspected pulmonary tuberculosis admitted between Aug 1, 2010, and July 31, 2013 (irrespective of the reason for admission), were prospectively investigated by culture, and by Xpert-MTB/RIF testing or smear microscopy, of tracheal aspirate samples. In the substudy, patients were randomly assigned (1:1), via a computer-generated allocation list, to smear microscopy or Xpert MTB/RIF. Participants, caregivers, and outcome assessors were not masked to group assignment. Only the laboratory staff were blinded to the clinical details of the participants. In November, 2012, Xpert MTB/RIF was adopted as the initial diagnostic test for respiratory samples in Western Cape province. Thereafter, patients received Xpert MTB/MIF and culture as standard of care. For the whole study cohort, the primary outcome was the frequency of bacteriologically confirmed tuberculosis. The primary endpoint of the randomised substudy was the proportion of culture-positive patients on treatment at 48 h after enrolment. The randomised substudy is registered with ClinicalTrials.gov, number NCT01530568. Findings: We investigated 341 patients for suspected pulmonary tuberculosis out of a total of 2309 ICU admissions. 46 (15%) of 317 patients included in the final analysis had a positive test for tuberculosis (Xpert MTB/RIF or culture). Culture-positive patients who failed to initiate treatment (adjusted HR 4·49, 95% CI 1·45-13·89) or who received inotropes (4·33, 1·49-12·60) were more likely to die. However, tuberculosis status was not associated with 28-day or 90-day mortality. In the substudy, we randomly assigned 115 patients to smear microscopy and 111 to Xpert MTB/RIF. Smear microscopy detected six (43%) of 14 culture-positive patients, and Xpert MTB/RIF detected 11 (100%) of 11 culture-positive patients (p=0·002). The proportion of culture-positive patients on treatment at 48 h was higher in the Xpert MTB/RIF group than in the smear microscopy group (11 [92%] of 12 vs nine [53%] of 17; p=0·043), although use of Xpert MTB/RIF had no effect on mortality or other patient outcomes. Interpretation: Tuberculosis is fairly common in ICUs in high-burden settings, and clinicians should screen and test patients for tuberculosis with Xpert MTB/RIF where available. This test improves diagnostic yield and rates of treatment initiation, and reduces unnecessary treatment, but might not increase the total number of patients on treatment when empirical treatment is widely used. A suspected diagnosis of pulmonary tuberculosis should not exclude patients from ICU care in resource-limited settings because mortality is unaffected by the presence of this disease. Funding: European and Developing Countries Clinical Trials Partnership, South African Medical Research Council, and the Discovery Foundation.

Artificial intelligence-based radiographic extent analysis to predict tuberculosis treatment outcomes: a multicenter cohort study

121. Improving Predictive Value of Phase 2a TB Drug Development Models Through PET/CT Imaging (nexgen EBA)

Machine learning algorithms to predict treatment success for patients with pulmonary tuberculosis

Can artificial intelligence (AI) be used to accurately detect tuberculosis (TB) from chest X-rays? An evaluation of five AI products for TB screening and triaging in a high TB burden setting

1951. Optimizing implementation of artificial intelligence-informed digital X-ray screening for tuberculosis through an age- and sex-differentiated approach

Chest X-ray Analysis With Deep Learning-Based Software as a Triage Test for Pulmonary Tuberculosis: An Individual Patient Data Meta-Analysis of Diagnostic Accuracy

Early user perspectives on using computer-aided detection software for interpreting chest X-ray images to enhance access and quality of care for persons with tuberculosis

A fully automatic artificial intelligence–based CT image analysis system for accurate detection, diagnosis, and quantitative severity evaluation of pulmonary tuberculosis

Evaluation of C-Reactive Protein and Computer-Aided Analysis of Chest X-rays as Tuberculosis Triage Tests at Health Facilities in Lesotho and South Africa

Using an Artificial Intelligence Approach to Predict the Adverse Effects and Prognosis of Tuberculosis

Evaluation of chest X-ray with automated interpretation algorithms for mass tuberculosis screening in prisons: a cross-sectional study

The utility of artificial intelligence in identifying radiological evidence of lung cancer and pulmonary tuberculosis in a high-burden tuberculosis setting

Computer-aided detection of tuberculosis from chest radiographs in a tuberculosis prevalence survey in South Africa: external validation and modelled impacts of commercially available artificial intelligence software

Expanding molecular diagnostic coverage for tuberculosis by combining computer-aided chest radiography and sputum specimen pooling: a modeling study from four high-burden countries

Usage of artificial intelligence tools in community-level X-ray triaging for tuberculosis in Chennai, Tamil Nadu

Triage of Persons With Tuberculosis Symptoms Using Artificial Intelligence-Based Chest Radiograph Interpretation: A Cost-Effectiveness Analysis

Harnessing Big Data to Optimize an Algorithm for Rapid Diagnosis of Pulmonary Tuberculosis in a Real-World Setting

Radiologist observations of computed tomography (CT) images predict treatment outcome in TB Portals, a real-world database of tuberculosis (TB) cases

Percent of lung involved in disease on chest X-ray predicts unfavorable treatment outcome in pulmonary tuberculosis

Identifying Predictors of Unfavorable Treatment Outcomes in Tuberculosis Patients