Macronutrient and micronutrient deficiencies are associated with tuberculosis (TB) incidence. However, evidence is limited on the impact of micronutrient (vitamins and minerals) supplementation among underweight individuals. We conducted a secondary data analysis of a randomised controlled trial of lipid nutritional supplements with and without high-dose vitamin and mineral supplementation (LNS-VM vs LNS) for underweight (Body Mass Index [BMI] <18.5 kg/m2) adults with human immunodeficiency virus (HIV) initiating antiretroviral therapy (ART) in Tanzania and Zambia (2011-2013). Incident TB disease diagnoses were extracted from trial records. We used multivariable Cox regression to estimate hazard ratios (HR) for the impact of receiving LNS-VM on TB incidence, and the dose-response relationship between baseline BMI and TB incidence. Overall, 263 (17%) of 1506 participants developed TB disease. After adjusting for age, sex, CD4 count, haemoglobin, and C-reactive protein, receiving LNS-VM was not associated with TB incidence (aHR [95%CI] = 0.93 [0.72-1.20]; p = 0.57) compared to LNS alone. There was strong evidence for an association between lower BMI and incident TB (aHR [95%CI]: 16-16.9kg/m2 = 1.15 [0.82-1.62] and <16kg/m2 = 1.70 [1.26-2.30] compared to 17-18.5kg/m2; linear trend p<0.01). There was strong evidence that the rate of developing TB was lower after initiating ART (p<0.01). In conclusion, the addition of micronutrient supplementation to LNS was not associated with lower TB incidence in this underweight ART-naive population.