• Infection control
• tuberculosis

In support of the urgent need for improvements to new entrant TB screening,1 which must encourage the diagnosis of both active and latent forms of TB, we would like to offer two audits of new entrant screening from an area with a low TB incidence (4.3/100 000).2

In 2006, we audited 29 new entrant referrals, all of whom had a chest x-ray reported by the Port Health Control Unit at Heathrow Airport as ‘abnormal’ (predominantly hilar calcification).3 Of the 29 referrals, 22 attended for local screening. Each received a tuberculin skin test (TST) and a repeat chest x-ray that was reported by a respiratory consultant and then by a consultant radiologist. Sixteen (73%) were subsequently reported as having a normal chest x-ray (and negative TST).

While the practical difficulties of screening large numbers of new entrants at the point of entry (in a short space of time) are high, inaccurate reporting of chest x-rays results in wasted resource and a financial burden that is passed on to both the new entrant and local TB services through the need for repeated screening.

Further, the NICE new entrant TB screening guidelines (2006)4 allow certain groups of new entrants to be screened solely via chest x-ray (CXR), limiting a TST to all those aged 0–15 and those aged 16–34 from sub-Saharan Africa. As the authors highlight, this potentially under-diagnoses the latent TB infection (LTBI).

To investigate this, we undertook a retrospective case-note analysis of 547 new entrants over a 44-month period (2006–2009).5 All patients were invited for screening using a locally adapted ‘Dorset’ algorithm that combined CXR and TST unless contraindicated. Each case was then re-evaluated using the NICE algorithm. This allowed direct comparison of each algorithm's ability to detect LTBI. Results: 397 (72%) new entrants attended screening, 41 (10.3%) patients were diagnosed with LTBI (all HIV negative). Comparison of the algorithms showed that only 27/41 cases (65.8 %) were detected when using the NICE algorithm. This represents a 34.1% shortfall in LTBI detection when following NICE guidance (95% CI 19.63% to 48.67%, 99% CI 15.04% to 53.26%).

The results from these two audits lend strength to the authors' argument that over-reliance on CXR alone is inadequate; combination screening with TST or IGRA should be considered. There remains a need for a robust national screening strategy that promotes the detection of latent as well as active tuberculosis.