Abstract
BACKGROUND: Treatment failure remains a major challenge in tuberculosis (TB) management. Rapid and objective assessment of treatment response is essential, as existing tools have limited accuracy and slow turnaround times. The PATHFAST TB LAM Ag assay (PATHFAST-LAM), an automated chemiluminescent enzyme immunoassay, was developed to quantify lipoarabinomannan (LAM) in sputum within 1 hour. Previous studies have shown a strong correlation between sputum LAM concentration and culture-based bacterial load. However, its clinical utility for predicting poor outcomes during treatment has not been prospectively evaluated. METHODS AND ANALYSIS: We will conduct a prospective longitudinal study enrolling newly diagnosed, bacteriologically confirmed patients with pulmonary TB at Rhodes Chest Clinic and Mbagathi County Referral Hospital in Nairobi, Kenya. We will follow participants throughout the 6-month treatment course, attempting to collect sputum weekly during weeks 1-4, biweekly during weeks 5-12 and monthly during months 3-6. We will measure LAM concentrations at these time points using the PATHFAST-LAM assay. The primary outcome is to assess whether changes in sputum LAM concentration during the intensive phase (baseline to week 4 and/or week 8) predict a composite poor outcome, defined as positive sputum culture at month 6, treatment failure, death during treatment or relapse within 3 months after treatment completion. The primary endpoint is the area under the curve from the receiver operating characteristic analysis, representing the predictive performance of changes in sputum LAM concentration for the composite poor outcome. We will identify the optimal cut-off value for LAM change and estimate sensitivity and specificity with 95% CIs using 2×2 tables. We will apply an adaptive design that allows sample-size re-estimation after interim analysis. ETHICS AND DISSEMINATION: The study was approved by the Kenya Medical Research Institute (KEMRI/SERU/CRDR/124/5241) and Nagasaki University (250619327). Findings will be disseminated through peer-reviewed publications and scientific meetings. TRIAL REGISTRATION NUMBER: NCT07157904.