Comparison of an interferon-gamma assay with tuberculin skin test for the diagnosis of tuberculosis infection prior to anti-tumour necrosis factor therapy

Abstract number: P1956

Casas S., Juanola X., Bordas X., Andreu A., Moure R., Alcaide F., Anibarro L., Cuchí E.M., Surís X., Guerrero M.R., Esteve M., Salvador G., Ortiz V., Martínez-Lacasa J., Cuquet J., Santín M.

Objectives: To compare the tuberculin skin test (TST) with a whole-blood interferon-gamma (IFN-g) assay in the diagnosis of tuberculosis infection (TBI) in patients with inflammatory diseases due to star anti-tumour necrosis factor (TNF) treatment.

Methods: A multicentre, cross-sectional study of the patients evaluated at 4 Spanish hospitals, from October 2006 to October 2007, before starting anti-TNF therapy. Diagnosis of TBI was based on TB exposure, chest X-ray, two-step TST and IFN-g assay (QuantiFERON® TB Gold-in Tube, QFT). According to the manufacturer's instruction, QFT was considered as positive when the TB antigen minus negative control IFN-g production geqslant R: gt-or-equal, slanted0.35 UI/mL; an indeterminate result was defined as either a positive control IFN-g response of leqslant R: less-than-or-eq, slant0.5 UI/mL or a negative control IFN-g level of >8 UI/mL. A positive TST was defined as a geqslant R: gt-or-equal, slanted5 mm induration. Agreement between TST and QFT was assessed by the Cohen kappa (k) index.

Results: 142 patients were screened for TBI, 51% women, with a mean age of 49 years. 44% had rheumatoid arthritis, 20% spondyloarthropathy, 19% cutaneous psoriasis, 10% psoriatic arthritis, 6% inflammatory bowel disease and 1% Behçet disease. 69% were under immunosuppressive treatment, mainly corticoids (39%) and methotrexate (38%). 23% were BCG vaccinated, 4% had a previous story of TBI and 6% had an abnormal X-ray. 33% of patients had a positive TST compared with 23% positive QFT (p < 0.01). Patients with some TB risk factor had 55% positive TST (p = 0.03) but only 32% positive QFT (p = 0.33). Overall agreement between the two tests was 81% (113/139), (k=0.54, 95% CI 0.38–0.7). Agreement was lower in patients with positive result by one of two tests (50%, p = 0.04). Agreement between the two tests was not influenced by BCG status (72% versus 83%, p = 0.16) neither by immunosuppressive treatment (81% versus 79%; p = 0.7). There were 3 (2%) indeterminate QFT results due to low IFN-g production in positive control.

Conclusions: Overall agreement between TST and QFT was fair, but among those patients with positive result by either of two tests was low. These results question the use of QFT as the only diagnostic test for TBI in this population. The rate of indeterminate results was low.

Session Details

Date: 19/04/2008
Time: 00:00-00:00
Session name: 18th European Congress of Clinical Microbiology and Infectious Diseases
Location: Barcelona, Spain
Presentation type:
Back to top