Evaluation of non-sputum-based diagnostics for pediatric tuberculosis: the Pediatric TB Diagnostic (PDTBDx) cohort protocol

This paper outlines the protocol for the Pediatric TB Diagnostic (PDTBDx) cohort study, a prospective observational trial in Nairobi, Kenya, designed to evaluate non-sputum-based diagnostics and novel biomarkers for improving the diagnosis and treatment monitoring of tuberculosis in children and adolescents.

The Gist

Imagine trying to find a tiny, invisible thief hiding inside a house. That's what doctors face when they try to diagnose Tuberculosis (TB) in children. The thief (the bacteria) is often hiding in very small numbers, and the house (the child's lungs) is hard to search because kids can't always cough up the "evidence" (sputum) that adults can.

This paper describes a new, massive detective mission called the PDTBDx study. Here is the story of how they plan to solve the mystery, explained simply.

The Problem: The "Silent Thief" in the House

TB is a deadly disease that kills thousands of children every year. The biggest issue? We often miss it.

• The Hard-to-Get Clue: To catch TB, doctors usually need a sample from the lungs. But asking a 3-year-old to cough up a deep lung sample is like asking them to pull a specific thread out of a tangled sweater while they are running around. It's hard, and often impossible.
• The Result: Because we can't get the sample, we often guess if a child has TB based on symptoms like a cough or fever. Sometimes we guess wrong, leading to delayed treatment and, sadly, death.

The Solution: A New "Super-Detective" Team

The researchers in Nairobi, Kenya, are launching a 24-month study to test new, easier ways to find this thief without needing that hard-to-get lung sample. They want to see if they can find the thief using blood or urine instead.

Think of this study as building a giant, high-tech library of clues.

1. The Recruitment: Casting a Wide Net

They are inviting over 400 children (under 15 years old) who are sick with symptoms that might be TB.

• Where? They are looking in hospitals, clinics, and even community centers in Nairobi.
• Who? They want a mix of kids: some who are very sick, some who are just a little sick, some with HIV, and some without. This ensures their "detective tools" work on all types of children, not just the easy cases.

2. The Process: The "24-Month Movie"

Instead of just taking one picture and leaving, this study is like filming a 24-month documentary of each child's health.

• The Start (Baseline): When a child joins, they get a full check-up. Doctors take X-rays, test their blood, and collect urine. They also try to get lung samples (sputum) if possible, but they know they might not get them.
• The Sequel (Follow-ups): The children return for check-ups at weeks 2, and months 1, 2, 4, 6, 12, and 24.
• The Goal: By watching how the children get better (or worse) over two years, the researchers can look back and say, "Yes, this child definitely had TB," or "No, this child was fine." This creates a "Gold Standard" truth that they can use to test their new tools.

3. The New Tools: Finding the Thief in the "Trash"

The main goal is to test non-sputum diagnostics.

• The Old Way: Digging through the house (lungs) to find the thief.
• The New Way: Checking the trash (blood and urine) for the thief's footprints.
  • Urine: They are testing a special strip that looks for a specific "fingerprint" (LAM) in the urine.
  • Blood: They are testing new, high-tech methods (like CRISPR and exosome analysis) that can spot tiny pieces of the TB bacteria floating in the blood.

They are also testing if these new tools can tell if the treatment is working. If the "footprints" in the blood disappear after a month of medicine, we know the medicine is working!

4. The "Truth Squad"

Because there is no perfect test for TB in kids, the researchers have a special panel of experts (like a jury). They will look at all the evidence—the symptoms, the X-rays, the treatment response, and the lab results—to decide the final verdict: Did this child have TB or not?

Once they know the "truth," they can test their new blood and urine tests against it.

• Did the new test catch the thief when the old test missed it?
• Did the new test give a false alarm?

Why This Matters

If this study succeeds, it's like giving doctors a metal detector instead of a shovel.

• Faster Diagnosis: They can find TB in kids quickly using a simple blood or urine test.
• Less Stress: No more forcing kids to cough up lung samples.
• Saving Lives: Early detection means early treatment, which stops the "thief" from hurting the child or spreading to others.

The Bottom Line

This paper is the instruction manual for a massive, 2-year detective story. The researchers are gathering a huge collection of clues (blood, urine, X-rays) from hundreds of children to prove that we can catch TB in kids using easy, non-invasive tests. If they succeed, it will change how the world diagnoses and treats childhood TB forever.

Technical Summary

Based on the provided preprint, here is a detailed technical summary of the Pediatric TB Diagnostic (PDTBDx) cohort protocol.

1. Problem Statement

Pediatric tuberculosis (TB) remains a critical global health challenge, causing approximately 172,000 deaths annually in children and adolescents (with 80% occurring in those under 5 years). The primary obstacles to effective management include:

• Diagnostic Limitations: Children are often paucibacillary (low bacterial load) and present with extrapulmonary disease, making the collection of high-quality respiratory specimens (sputum) difficult.
• Reliance on Clinical Criteria: Due to the difficulty in obtaining microbiologic confirmation, diagnoses and treatment decisions are frequently based on imperfect clinical criteria rather than definitive microbiologic evidence.
• Lack of Validated Non-Sputum Diagnostics: There is an urgent need for non-sputum-based diagnostic tools (e.g., urine or blood-based) that are accurate for children. However, evaluating these new assays is complicated by the lack of a "perfect" gold standard for pediatric TB, necessitating well-characterized cohorts with rigorous follow-up to establish reference standards.

2. Methodology

The PDTBDx is a prospective, longitudinal observational cohort study designed to evaluate non-sputum-based diagnostics and treatment response in children.

• Study Design & Setting:

  • Location: Nairobi, Kenya. Recruitment occurs across inpatient wards, outpatient clinics, and community-based HIV/TB care facilities.
  • Population: Children ≤15 years of age with presumptive TB (symptoms such as persistent cough, fever, weight loss, night sweats).
  • Sample Size: Target enrollment of >400 participants.
  • Duration: Follow-up for 24 months.
  • Timeline: Enrollment began in May 2022.

• Data Collection & Procedures:

  • Baseline Visit: Comprehensive assessment including symptom questionnaires, physical exams, anthropometry, HIV testing (and CD4/viral load if positive), Tuberculin Skin Test (TST)/IGRA, and Chest X-ray (CXR) with standardized scoring.
  • Specimen Collection:
    • Respiratory: Spontaneous sputum, induced sputum (via hypertonic saline), or gastric aspirates. Tested via GeneXpert Ultra, microscopy, and culture.
    • Urine: Collected for Lipoarabinomannan (LAM) testing and future biomarker analysis.
    • Blood: Whole blood, plasma, and PBMCs collected for CBC, CRP, and storage for novel assays.
  • Follow-up: Visits at Week 2, and Months 1, 2, 4, 6, 12, and 24. These visits monitor treatment response, collect repeat urine/blood samples, and assess symptom resolution.

• Reference Standard & Classification:

  • Participants are classified using NIH consensus statement guidelines (Graham 2015), modified to include extrapulmonary TB.
  • A two-clinician review panel adjudicates disease status (Confirmed, Unconfirmed, or No TB) based on clinical presentation, microbiology, CXR, and treatment response.
  • Blinding: Clinicians and investigators are blinded to the results of novel diagnostic assays during treatment and classification to prevent bias.

• Novel Diagnostic Evaluation:

  • The study evaluates emerging biomarkers including exosome-based assays, CRISPR-based diagnostics, M. tuberculosis extracellular vesicles (Mtb-EVs), and cell-free DNA.
  • Statistical Analysis:
    • Primary metrics: Sensitivity, specificity, PPV, NPV, and AUC-ROC.
    • Bayesian Latent Class Analysis: Used to estimate the accuracy of new diagnostics in the absence of a perfect gold standard, accounting for conditional dependence between tests and adjusting for covariates (age, HIV status, malnutrition).

3. Key Contributions

• Protocol Harmonization: The study adopts principles from the RePORT initiative to create a standardized, harmonized protocol for pediatric TB cohorts globally, facilitating cross-study comparisons.
• Comprehensive Biospecimen Repository: Creates a large, longitudinal repository of blood and urine samples from children with confirmed, unconfirmed, and non-TB diagnoses, linked to detailed clinical and radiographic data.
• Rigorous Reference Standard: By utilizing a 24-month follow-up and a blinded adjudication panel, the study overcomes the "imperfect gold standard" problem, allowing for more accurate validation of new diagnostic tools.
• Inclusion of Unconfirmed Cases: Unlike many studies that exclude children without microbiologic confirmation, PDTBDx includes them, reflecting real-world epidemiology and improving the generalizability of diagnostic performance estimates.

4. Results

• Note: As this is a study protocol published prior to the completion of data analysis, no empirical results (e.g., sensitivity/specificity rates of the new assays) are reported in this document.
• The document outlines the planned analytical framework, including the use of Bayesian models to estimate diagnostic accuracy and the specific timelines for data collection.

5. Significance

• Advancing Pediatric TB Diagnostics: The study aims to validate non-invasive, non-sputum diagnostics that can be used in routine clinical settings, potentially reducing the time to diagnosis and mortality in children.
• Treatment Response Monitoring: The longitudinal design allows for the evaluation of biomarkers not just for diagnosis, but for monitoring treatment efficacy and detecting post-TB lung disease.
• Policy Impact: Findings are expected to guide the development of WHO and national guidelines for pediatric TB screening and diagnosis, particularly in high-burden, resource-limited settings like Sub-Saharan Africa.
• Scientific Infrastructure: The establishment of a centralized, standardized cohort in Kenya provides a critical platform for the rapid evaluation of emerging molecular and imaging technologies.