Background Despite a recent decline in the annual incidence of tuberculosis (TB) in the UK, rates remain higher than in most Western European countries. The detection and treatment of latent TB infection (LTBI) is an essential component of the UK TB control programme. Objectives To assess the prognostic value and cost-effectiveness of the current two interferon gamma release assays (IGRAs) compared with the standard tuberculin skin test (TST) for predicting active TB among untreated individuals at increased risk of TB: (1) contacts of active TB cases and (2) new entrants to the UK from high-TB-burden countries. Design A prospective cohort study and economic analysis. Participants and setting Participants were recruited in TB clinics, general practices and community settings. Contacts of active TB cases and migrants who were born in high-TB-burden countries arriving in the UK were eligible to take part if they were aged ≥ 16 years. Main outcome measures Outcomes include incidence rate ratios comparing the incidence of active TB in those participants with a positive test result and those with a negative test result for each assay, and combination of tests and the cost per quality-adjusted life-year (QALY) for each screening strategy. Results A total of 10,045 participants were recruited between May 2010 and July 2015. Among 9610 evaluable participants, 97 (1.0%) developed active TB. For the primary analysis, all test data were available for 6380 participants, with 77 participants developing active TB. A positive result for TST^a (positive if induration is ≥ 5 mm) was a significantly poorer predictor of progression to active TB than a positive result for any of the other tests. Compared with TST^b [positive if induration is ≥ 6 mm without prior bacillus Calmette–Guérin (BCG) alone, T-SPOT^®.TB (Oxford Immunotec Ltd, Oxford, UK), TST^a + T-SPOT.TB, TST^a + IGRA and the three combination strategies including TST^b were significantly superior predictors of progression. Compared with the T-SPOT.TB test alone, TST^a + T-SPOT.TB, TST^b + QuantiFERON^® TB Gold In-Tube (QFT-GIT; QIAGEN GmbH, Hilden, Germany) and TST^b + IGRA were significantly superior predictors of progression and, compared with QFT-GIT alone, T-SPOT.TB, TST^a + T-SPOT.TB, TST^a + QFT-GIT, TST^a + IGRA, TST^b + T-SPOT.TB, TST^b + QFT-GIT and TST^b + IGRA were significantly superior predictors of progression. When evaluating the negative predictive performance of tests and strategies, negative results for TST^a + QFT-GIT were significantly poorer predictors of non-progression than negative results for TST^a, T-SPOT.TB and TST^a + IGRA. The most cost-effective LTBI testing strategies are the dual-testing strategies. The cost and QALY differences between the LTBI testing strategies were small; in particular, QFT-GIT, TST^b + T-SPOT.TB and TST^b + QFT-GIT had very similar incremental net benefit estimates. Conclusion This study found modest differences between tests, or combinations of tests, in identifying individuals who would go on to develop active TB. However, a two-step approach that combined TST^b with an IGRA was the most cost-effective testing option. Implications for practice and future research The two-step TST^b strategy, which stratified the TST by prior BCG vaccination followed by an IGRA, was the most cost-effective approach. The limited ability of current tests to predict who will progress limits the clinical utility of tests. The implications of these results for the NHS England/Public Health England national TB screening programme for migrants should be investigated. Study registration This study is registered as NCT01162265. Funding The National Institute for Health Research Health Technology Assessment programme.