Direct microscopy versus sputum cytology analysis and bleach sedimentation for diagnosis of tuberculosis: a prospective diagnostic study

BACKGROUND Diagnostic options for pulmonary tuberculosis in resource-poor settings are commonly limited to smear microscopy. We investigated whether bleach concentration by sedimentation and sputum cytology analysis (SCA) increased the positivity rate of smear microscopy for smear-positive tuberculosis. METHODS We did a prospective diagnostic study in a Médecins Sans Frontières-supported hospital in Mindouli, Republic of Congo. Three sputum samples were obtained from 280 consecutive pulmonary tuberculosis suspects, and were processed according to WHO guidelines for direct smear microscopy. The remainder of each sputum sample was homogenised with 2.6% bleach, sedimented overnight, smeared, and examined blinded to the direct smear result for acid-fast bacilli (AFB). All direct smears were assessed for quality by SCA. If a patient produced fewer than three good-quality sputum samples, further samples were requested. Sediment smear examination was performed independently of SCA result on the corresponding direct smear. Positivity rates were compared using McNemar's test. RESULTS Excluding SCA, 43.2% of all patients were diagnosed as positive on direct microscopy of up to three samples. 47.9% were diagnosed on sediment microscopy, with 48.2% being diagnosed on direct microscopy, sediment microscopy, or both. The positivity rate increased from 43.2% to 47.9% with a case definition of one positive smear (≥1 AFB/100 high power fields) of three, and from 42.1% to 43.9% with two positive smears. SCA resulted in 87.9% of patients producing at least two good-quality sputum samples, with 75.7% producing three or more. Using a case definition of one positive smear, the incremental yield of bleach sedimentation was 14/121, or 11.6% (95% CI 6.5-18.6, p = 0.001) and in combination with SCA was 15/121, or 12.4% (95% CI 7.1-19.6, p = 0.002). Incremental yields with two positive smears were 5/118, or 4.2% (95% CI 1.4-9.6, p = 0.062) and 7/118, or 5.9% (95% CI 2.4-11.8, p = 0.016), respectively. CONCLUSIONS The combination of bleach sedimentation and SCA resulted in significantly increased microscopy positivity rates with a case definition of either one or two positive smears. Implementation of bleach sedimentation led to a significant increase in the diagnosis of smear-positive patients. Implementation of SCA did not result in significantly increased diagnosis of tuberculosis, but did result in improved sample quality. Requesting extra sputum samples based on SCA results, combined with bleach sedimentation, could significantly increase the detection of smear-positive patients if routinely implemented in resource-limited settings where gold standard techniques are not available. We recommend that a pilot phase is undertaken before routine implementation to determine the impact in a particular context.