Link
Our recommendations for screening are a blend of those in the published guidelines: screening is recommended only for population subgroups with a high prevalence of latent tuberculosis infection or those with a high likelihood of progression from latent tuberculosis infection to active disease (Tables 1 and 2). Although data informing the optimal approach to screening in various subgroups are limited, we prefer the use of IGRAs when the prevalence of recent infection is likely to be high: these populations include close contacts of patients with active tuberculosis, recently arrived foreign-born persons, drug users, incarcerated persons, and homeless persons. IGRAs are also of value for screening persons who have received the BCG vaccine. For populations in which the prevalence is low or infection is likely to have been remote, we prefer the tuberculin skin test because IGRAs have either not been well studied (e.g., among smokers or persons with diabetes mellitus) or have performed poorly (e.g., among health care workers). Screening with either test is not recommended in persons at low risk, since such persons are much more likely to have false positive than true positive results.The CDC clarify the recommendations against testing low risk persons; it is not correct to assume that IGRA have the same rate of false diagnosis as the skin test;
As with the TST, IGRAs generally should not be used for testing persons who have a low risk for both infection and progression to active tuberculosis if infected (except for those likely to be at increased risk in the future). Screening such persons diverts resources from higher priority activities and increases the number of false-positive results. Even with a test specificity approaching 99%, when the prevalence of M. tuberculosis infection is ≤1%, the majority of positive results will be false positives. If persons at low risk for both infection and progression are to be tested, selection of the test with the greatest specificity will minimize false-positive results, reduce unnecessary evaluation and treatment, and minimize the potential for adverse events from unnecessary treatment.Obviously when it comes to testing low risk it would be logical to use the test with the highest specificity.
