Evaluation of FASTPlaqueTB to diagnose smear-negative tuberculosis in a peripheral clinic in Kenya.

OBJECTIVE To evaluate the performance and feasibility of FASTPlaqueTB in smear-negative tuberculosis (TB) suspects in a peripheral clinic after laboratory upgrading. DESIGN Patients with cough > or=2 weeks, two sputum smear-negative results, no response to 1 week of amoxicillin and abnormal chest X-ray were defined as smear-negative suspects. One sputum sample was collected, decontaminated and divided into two: half was tested with FASTPlaqueTB in the clinic laboratory and the other half was cultured on Löwenstein-Jensen medium in the Kenyan Medical Research Institute. Test sensitivity and specificity were evaluated in all patients and in human immunodeficiency virus (HIV) infected patients. Feasibility was assessed by the contamination rate and the resources required to upgrade the laboratory. RESULTS Of 208 patients included in the study, 56.2% were HIV-infected. Of 203 FASTPlaqueTB tests, 95 (46.8%) were contaminated, which interfered with result interpretation and led to the interruption of the study. Sensitivity and specificity were respectively 31.2% (95%CI 12.1-58.5) and 94.9% (95%CI 86.8-98.4) in all patients and 33.3% (95%CI 9.9-65.1) and 93.9% (95%CI 83.1-98.7) in HIV-infected patients. Upgrading the laboratory cost euro 20,000. CONCLUSION FASTPlaqueTB did not perform satisfactorily in this setting. If contamination can be reduced, in addition to laboratory upgrading, its introduction in peripheral clinics would require further assessment in smear-negative and HIV co-infected patients and test adaptation for friendlier use.