Development of a Diagnostic Prediction Score for Tuberculosis in Hospitalized Children With Severe Acute Malnutrition

There is now strong evidence that undiagnosed and untreated TB increases the risk of death in children, especially those severely malnourished who are highly vulnerable. Specific decision-making tools are therefore urgently needed to guide clinicians from high TB burden and low-income countries to initiate treatment quickly in children with SAM with suspected TB.

A diagnostic prediction score and algorithm was recently proposed by the investigators for TB treatment decision in HIV-infected children with presumptive TB (developed in the ANRS 12229 PAANTHER 01 study). Based on easily collected clinical features, chest X-Ray (CXR), Xpert MTB/RIF, and abdominal ultrasonography, the score aims to help clinicians make a same-day treatment decision. Such a prediction score improving TB diagnosis and shortening time to treatment initiation would be a key benefit in children with SAM.

Based on this experience, the investigators are proposing a diagnostic cohort study enrolling hospitalized severely malnourished children. The study will include the evaluation of several diagnostic tests that could be integrated in the development of a prediction model and subsequent score for the diagnosis of TB in hospitalized children with SAM. This will include Xpert MTB/RIF Ultra performed on one nasopharyngeal aspirate (NPA) and one stool sample, CXR, Quantiferon (QFT) Interferon-Gamma Release Assay (IGRA), Monocyte-to-lymphocyte ratio (MLR), and ultrasonography, which has shown its interest for the diagnosis of TB in both HIV-infected adults and children. In the PAANTHER study, it detected abdominal lymphadenopathy in 50% of culture confirmed TB cases and 35% of all confirmed and unconfirmed cases, with a specificity of 85%.

Using logistic regression, a score will be developed for TB diagnosis, considering confirmed and unconfirmed TB as reference diagnosis, in hospitalized children with SAM. As a secondary objective, and in order to reduce costs, sample collection, and complexity of the diagnostic process, a first-step screening score (excluding Ultra, abdominal ultrasound, and CXR if possible) will be developed to identify children with presumptive TB who would benefit from further diagnostic testing.

Both scores will be internally validated using resampling and will be incorporated in a stepwise algorithm to guide practical implementation of the screening and diagnosis process. The stepwise algorithm will be discussed with local clinicians involved in the study to better adapt it for future use in their routine practice.