Objective: To compare the efficacy of the conventional skin test and a novel flow cytometric whole blood assay in the diagnosis of latent tuberculosis infection (LTBI) in patients with rheumatological diseases evaluated for treatment with TNF-alpha-blocking agents.
Methods: Prospective study of 97 consecutively enrolled patients, who were assessed for the presence of LTBI through clinical history, Mendel-Mantoux skin testing and chest X-ray. In addition, T-cell reactivity towards tuberculin (PPD, purified protein derivative) and the Mycobacterium tuberculosis-specific proteins ESAT-6 and CFP-10 was determined ex vivo using a flow cytometric whole blood assay.
Results: After standard screening, 15% of patients receiving TNF-alpha-blocking therapy were pretreated with isoniazide (INH), another 5% of patients did not receive TNF-alpha-blocking therapy because of LTBI. PPD-reactivity in the skin was observed in 14% of patients compared with 39% with the whole blood test. Analysis of the M. tuberculosis-specific response to ESAT-6 and CFP-10 revealed positive results in 16% of patients. Using a decision tree incorporating history, chest X-ray and either skin-test or ESAT-6/CFP-10 results, 18 or 22% of the patients, respectively, were classified as latently infected with M. tuberculosis. Four patients treated with INH because of a positive skin reaction did not show reactivity to ESAT-6/CFP-10 in the whole blood assays. Another six patients not pretreated with INH because of negative skin tests would have received INH, had the results of the whole blood assay been taken into account.
Conclusion: The Mendel-Mantoux skin test has a low sensitivity and specificity for the diagnosis of LTBI in this cohort of patients, potentially resulting in both over- and under-treatment with prophylactic INH when compared with the flow cytometric analysis of whole blood T-cell reactivity to proteins specific to M. tuberculosis. Use of T-cell based in vitro tests may help to refine diagnostic testing for LTBI.