Evaluating Diagnostics for Paediatric Tuberculosis by Blood Culture

• Diagnostic yield of TB culture [ Time Frame: At baseline - day 1 of study ]
  Number of positive TB cultures versus number of positive TB direct smears for expectorate and/or gastric aspirate.
  
  
• Diagnostic yield of TB culture [ Time Frame: At baseline - day 1 of study ]
  Number of positive TB cultures versus number of positive MODS cultures for expectorate and/or gastric aspirate.
  
  
• Diagnostic yield of urine versus expectorate or gastric aspirate for TB culture [ Time Frame: At baseline - day 1 of study ]
  Number of positive TB cultures in urine versus expectorate and/or gastric aspirate.
  
  

Tuberculosis (TB) is a major cause of morbidity and mortality among children in developing nations. Symptom-based diagnostic criteria are non-specific and culture confirmation is challenging, as sputum samples are often believed to be to too cumbersome to obtain from small children and specimens typically have low yield due to the paucibacillary nature of pediatric TB. Culture confirmation may be obtained in as few as 10% of cases of suspected pediatric TB. For these reasons, the true extent of the (drug-resistant) TB epidemic in children is unknown. Thus, either clinicians begin empiric treatment without diagnosis or no treatment is given at all. Current laboratory methods, if available at all in resource poor settings, employ smears from expectorated sputa or gastric aspirates which have low sensitivity in children. While more rapid diagnostic techniques such as PCR based tests have been developed, there is still poor sensitivity in children. Improving the diagnosis of pediatric TB must focus on better efforts, including more aggressive strategies to uncover disseminated disease.

Culture confirmation of disseminated disease can be obtained from blood, urine, cerebrospinal fluid (CSF), peritoneal and pleural fluid, or purulent material from lymph node aspirates, abscesses or otorrhea. Unfortunately, little is known about the overall yield from these various specimens in children. From pilot data collected among children at NHP, we know that it is feasible to collect and test various bodily fluid specimens for TB culture.

Although WHO guidelines encourage body fluid collection in order to make a diagnosis of TB in children, at present in NHP, blood and urine cultures are not obtained for mycobacterial culture. However, this study seeks to demonstrate that routine investigation of blood and urine will augment the yield of traditional sputum culture for children in whom disseminated disease is more likely. Improved culture confirmation will allow DST and a more accurate description of the drug-resistant TB epidemic for children in the region.