Gross appearance of sputum and factors associated with genexpert outcome among smear-negative adult presumptive tuberculosis patients at Mulago Hospital Kampala.

Abstract

Background: The introduction of GeneXpert in the diagnostic algorithm of TB was envisaged to cause improvements in the diagnosis. Although some factors that influence its performance including HIV status, sputum volume have been identified, it is still not clear how sputum gross appearance influences its diagnostic performance. The findings from this study may influence health policy to strengthen the laboratory guidelines and SOPs regarding PTB diagnosis by GeneXpert. Objective: To determine the effect of gross appearance of sputum and factors on GeneXpert outcome among smear negative adult presumptive TB patients at Mulago Hospital in Kampala. Methods: A retrospective review of records of inpatients undergoing TB evaluation at Mulago Hospital Complex Kampala under MIND (Mulago In-patient Non-invasive Diagnosis of opportunistic pneumonias). The records were collected from August 2009 to December 2014. Socio-demographic, clinical and laboratory factors in addition to gross appearance of sputum were recorded. Operating characteristics were summarized in form of proportions and ratios i.e sensitivity, specificity, positive and negative predictive values and likelihood ratios using STATA version 12. Also 95% CIs were constructed for these statistics for estimation purposes. Socio-demographic and clinical factors were summarized as medians and proportions, and their effect on a true GeneXpert outcome was obtained as prevalence ratios using Robust Poisson regression analysis. Results: Among 1321 records of adult Pulmonary TB suspects, 924 records were analyzed and 794 records (155 true positive Xpert result and 639 true negative Xpert result) were analyzed with Robust Poisson regression analysis. Salivary sample had statistically similar sensitivity with purulent and mucoid of 69.8% (95%CI 53.9-82.8), 61.5% (95%CI 31.6-86.1) and 58.9% (95%CI 51.5-65.9) respectively. Blood-stained sputum had lowest sensitivity of 28.6% (95%CI 8.39-58.1). HIV sero-positivity and blood-stained sputum were found associated with true GeneXpert positivity. The adjusted prevalence ratio (aPR) for HIV sero-positivity was 11.0 (95%CI 4.49-27.0) and blood-stained sputum vs salivary possessed aPR of 0.3 (95%CI 0.1-0.8). xiii

Conclusions: GeneXpert testing demonstrated similar sensitivity for the three sputum specimens; salivary, purulent and mucoid sputum among smear-negative individuals. However, GeneXpert sensitivity on blood stained sputum was significantly low, suggesting it may be an unsuitable sample for diagnostic testing. Human Immune Virus (HIV) positive individuals are more likely to yield a true GeneXpert positive result.