Implementation of ML Flow for leprosy contacts in Brazil: Opportunities, pitfalls, and safeguards

Brazil has the second-highest leprosy burden worldwide, with approximately 20,000 new cases reported annually, many diagnosed with advanced disease and disability. To support earlier detection, the Ministry of Health recently approved the ML Flow rapid test for contact evaluation. ML Flow detects immunoglobulin M antibodies against phenolic glycolipid-I of Mycobacterium leprae, can be performed at the point of care using finger-prick blood, and yields results within minutes. ML Flow offers important operational advantages. It enables same-visit counseling and may support risk-stratified household-contact follow-up. However, its value is context-dependent. Because anti-phenolic glycolipid-I immunoglobulin M responses correlate with bacillary burden, positivity is more frequent in multibacillary disease, whereas many paucibacillary and pure neural cases have absent or low antibody levels. A seronegative result therefore does not exclude disease. Among asymptomatic contacts, seropositivity varies widely across studies, and in previously treated individuals antibodies may remain detectable for years. Available evidence suggests that seropositive contacts are at increased risk of incident leprosy, although predictive performance varies across settings. Serology alone is not diagnostic, and misuse may lead to stigma, anxiety, unnecessary referrals, and diversion of resources. The contribution of ML Flow is therefore implementation-dependent. In settings with standardized counseling, scheduled re-examination, and reliable referral pathways, it may support risk-stratified follow-up. In settings where these elements are weak, benefits may be limited. Brazil offers an important programmatic setting in which to evaluate this strategy, but only with safeguards: integration with dermato-neurological examination, clear protocols stating that seropositivity is not diagnostic, structured follow-up pathways, quality-controlled training, and systematic recording in the Brazilian Unified Health System information systems. Under these conditions, ML Flow may contribute to earlier diagnosis and disability reduction; without them, it risks adding workload without improving care.