Background Two commercial blood assays for the diagnosis of latent tuberculosis infection—T-SPOT.TB and QuantiFERON-TB Gold—have been separately compared with the tuberculin skin test. Our aim was to compare the efficacy of all three tests in the same population sample. Methods We did a prospective study in 393 consecutively enrolled patients who were tested simultaneously with T-SPOT.TB and QuantiFERON-TB Gold because of suspected latent or active tuberculosis. 318 patients also had results available for a tuberculin skin test. Findings Overall agreement with the skin test was similar (T-SPOT.TB κ=0·508, QuantiFERON-TB Gold κ=0·460), but fewer BCG-vaccinated individuals were identified as positive by the two blood assays than by the tuberculin skin test (p=0·003 for T-SPOT.TB and p<0.0001 for QuantiFERON-TB Gold). Indeterminate results were significantly more frequent with QuantiFERON-TB Gold (11%, 43 of 383) than with T-SPOT.TB (3%, 12 of 383; p<0·0001) and were associated with immunosuppressive treatments for both tests. Age younger than 5 years was significantly associated with indeterminate results with QuantiFERON-TB Gold (p=0.003), but not with T-SPOT.TB. Overall, T-SPOT.TB produced significantly more positive results (38%, n=144, vs 26%, n=100, with QuantiFERON-TB Gold; p<0·0001), and close contacts of patients with active tuberculosis were more likely to be positive with T-SPOT.TB than with QuantiFERON-TB Gold (p=0·0010). Interpretation T-SPOT.TB and QuantiFERON-TB Gold have higher specificity than the tuberculin skin test. Rates of indeterminate and positive results, however, differ between the blood tests, suggesting that they might provide different results in routine clinical practice.