Dynamic, quantitative ESAT6-CFP10 skin test for tuberculosis risk prediction: a large-scale, multi-center,...

Area of Science:

• Infectious Diseases
• Public Health
• Diagnostic Testing

Background:

• Predicting progression from Mycobacterium tuberculosis (M. tb) infection to active disease is a significant public health challenge.
• Current risk stratification methods require improvement for effective targeted prevention strategies.

Purpose of the Study:

• To evaluate a serial, quantitative ESAT6-CFP10 (EC) antigen-based skin test strategy for improved tuberculosis risk stratification.
• To assess the effectiveness of this strategy in a large-scale prospective cohort of close contacts.

Main Methods:

• Enrolled 73,761 close contacts during school-based tuberculosis outbreaks.
• Utilized baseline and serial (8-12 weeks) EC skin testing, chest radiography, and symptom screening.
• Compared single versus serial test strategies using Cox models, ROC, and PR curves for incident tuberculosis identification.

Main Results:

• The serial EC test strategy (C-statistic: 0.806) significantly outperformed a single test (C-statistic: 0.722).
• EC response size demonstrated a dose-dependent association with incident tuberculosis risk (7% higher hazard per 1-mm increase).
• The combined strategy achieved 65.0% sensitivity and 96.1% specificity at a ≥5 mm threshold.

Conclusions:

• A "test twice, take maximum" EC strategy offers superior risk stratification for tuberculosis prevention.
• This pragmatic approach enables efficient targeting of preventive treatment, especially in resource-constrained settings.
• Quantitative EC skin testing provides a practical alternative for programmatic risk stratification in tuberculosis control.