During the fall of 2001, a tuberculosis outbreak caused by Mycobacterium bovis occurred in a conditioned colony of rhesus (Macaca mulatta) and cynomolgus (Macaca fascicularis) macaques at Stanford University School of Medicine. During this outbreak, we evaluated the diagnostic performance of a new in vitro tuberculosis screening test (PRIMAGAM). The PRIMAGAM test measures the interferon-gamma (IFNgamma) response to purified protein derivatives (PPDs) of M. bovis and M. avium. On the basis of the results of the last test administered before necropsy, the PRIMAGAM test had good sensitivity (68%) and excellent specificity (97%), compared with the disease status, as determined by the presence or absence of gross and/or histologic lesions indicative of tuberculosis. By contrast, sensitivity and specificity of the tuberculin skin test (TST) was 84 and 87%, respectively. Both tests suffered from intermittent positive and negative reactions on repeat testing. Overall, however, there was no significant difference (P = 0.09, McNemar's chi2-test) and moderate agreement (kappa = 0.52) between these two tests. Lastly, the IFNgamma response to bovine PPD was significantly lower in infected cynomolgus macaques. Moreover, each test failed to detect tuberculosis in three cynomolgus macaques. Fortunately, they were different animals; therefore, we recommend the parallel use of the TST and PRIMAGAM test for maximal overall sensitivity in a tuberculosis screening program, especially for cynomolgus macaques.