Article Text
PDF
Abstract
Background Interferon-gamma release assays (IGRAs) are being used to confirm significant exposure to tuberculosis (TB) after a positive tuberculin skin test (TST). We explored the variability of a positive test in a large UK TB clinic amongst contacts of active disease.
Methods The records of all individuals notified with active TB and of their screened contacts from 2012–3 were reviewed retrospectively. Uni- and multivariate analysis identified variables associated with active TB and/or QuantiFERON® test positivity in contacts.
Results Of 203 notified tuberculosis index cases, 470 contacts were screened. 116 had immunological evidence of TB exposure including 11 with active disease. The estimated background rate of positive TB immune response was 18% and was the same in contacts of index cases who were later denotified (7/39) as in contacts of extrapulmonary disease (24/132).
On univariate analysis, index case variables associated with a positive TST and IGRA were a pulmonary vs. extrapulmonary site of disease (OR 2.1; 95% confidence interval 1.3–3.5) and, for contacts of pulmonary TB (PTB), presence of cough (2.5; 1.1–4.0), duration of cough (p = 0.04),sputum smear positivity (1.7; 1.0–2.8), sputum smear grade (p = 0.02), radiographic extent of disease (p = 0.008), and perhaps pulmonary cavities (1.6; 1.0–2.6). No effect was detected of gender, smoking status, or TB strain of the index case (Beijing or Cameroon vs. other lineage), or nature of contact (household vs. other), even when considering only child contacts.
Contact variables associated with positive tests were birth in a high TB prevalence country (2.3; 1.4–3.6), age ≥12 years (2.0; 1.2–3.3), smoking (2.2; 1.0–5.1) and perhaps alcohol excess (3.2; 0.9–10.7), but not diabetes or BCG vaccination. There were insufficient data to investigate an effect of HIV and homelessness. Contact country of birth (1.9; 1.1–3.1) and index site of disease (2.0; 1.2–3.4), sputum smear positivity (2.6; 1.3–5.6) and radiographic extent of disease (p = 0.008) remained significant on multi-variate analysis.
Conclusion The likelihood of significant TB exposure is a consequence of bacillary load, radiographic extent of disease and coughing behaviour.