SignificanceDespite appropriate antibiotic treatment, tuberculous meningitis carries a high mortality ascribed to overexuberant inflammation. Genetic variations in the enzyme, LTA4H, alter inflammation, with individuals carrying the inflammation-associatedLTA4Hvariant benefitting from antiinflammatory steroids administered alongside antibiotics. A prior study found poor correlation betweenLTA4Hgenotype and cerebrospinal fluid levels of cytokines, key mediators of inflammation. The study used “frequentist” statistical methods that can fail to detect true differences. Using Bayesian statistics, which can detect significant differences not found by frequentist methods, we found good correlation betweenLTA4Hgenotype and cytokine levels, and cytokine levels and outcome even independent ofLTA4Hgenotype. These findings suggest that LTA4H and additional inflammation factors affect outcome and suggest tailoring steroid therapy to cytokine levels.