Introduction Diagnosis of tuberculosis (TB) ideally involves culture and sensitivity of the organism but in low income countries this is not practised routinely. The World Health Organisation estimates that substandard detection occurs in 40% of patients globally, with many diagnosed on clinical suspicion or response to medication. TB produces a strong antibody response suitable for simple, inexpensive and rapid serodiagnostic assays. Ongoing evaluation of a new point of care rapid serological test based on lateral flow immunochromatography (TB-ST Rapid Test, Lionex, Germany) has shown 100% specificity, with no false positive tests in normal controls and latently infected patients, but sensitivity of 36%, with false negatives in culture proven TB.

Aim Using this test in a resource poor setting to investigate whether cases of active TB may be being missed by current diagnostic methods.

Methods 498 patients in chest and HIV clinics in two rural Kenyan hospitals were investigated with the TB-ST Rapid Test and a scored questionnaire to determine symptoms and risk of TB. Results were compared with clinical diagnoses made, usually based on symptoms alone. Chest radiographs were performed in only 111 and sputum smears in 75.

Results 127/498 patients were HIV positive. Of these, only 59(46%) had a clinical diagnosis of TB, whereas 87(68.5%) had significant TB symptoms and/or risk factors, and 82(64.6%) tested TB-ST positive (p<0.001). Therefore clinical diagnosis accounted for significantly fewer diagnoses of active TB than suggested by either symptom and risk score or TB-ST rapid results in the HIV+ population. Of the 375 HIV− patients, 73 (19.7%) had a clinical diagnosis of TB, 46 (12.3%) scored positive for TB on the questionnaire, and 149 (40.2%) were TB-ST+ (p<0.001). Abstract P165 Table 1 shows positive and negative TB-ST results related to positive and negative sputum smears in 34 HIV+ and 41 HIV− patients.

Conclusions Many more patients had positive TB-ST and risk and symptom scores than were being diagnosed with active TB, suggesting that TB may not always be being diagnosed or treated. Sputum smears were in greater agreement with the TB-ST in HIV− but not HIV+ patients, in whom there were considerably fewer positive smears than TB-ST results.