Following an extensive contact tracing exercise at a school in a London borough with one of highest tuberculosis (TB) rates in England, we estimated the background prevalence of latent TB infection to be significantly less than the widely accepted 10%. We screened 271 pupils aged 14–15 years in two groups: 96 pupils in group 1 had significant exposure (>8 h/week in the same room) to a case of infectious TB and 175 in group 2 who had minimal exposure. In group 1, 26% were diagnosed with latent or active TB, compared to 6.3% in group 2. Risk factors for TB infection (e.g. previous exposure or link to high-prevalence communities) were analysed using a cohort study design. In the univariable analysis only being in contact group 1 was statistically significantly associated with being a case (OR 5.25, 95%, P < 0.001). In the multivariable model contact group 1 remained significantly associated with being a case (adjusted OR 4.40, P = 0.001). We concluded that the 6.3% yield of TB infection in contact group 2 is either similar to or higher than the background prevalence rate of latent TB infection (LTBI) in this high TB prevalence London borough. Other parts of England with lower TB prevalence are likely to have even lower LTBI rates.