Rédiger un PREreview

Mycoplasma pneumoniae (MP) infects tens of millions of children in an epidemic year, yet registers as a rounding error in the mortality metrics by which global health allocates resources. We audited this invisibility by triangulating three evidence streams: six national and multinational surveillance systems; a M. pneumoniae–specific secondary analysis of Global Burden of Disease (GBD) 2023 data; and GBD-derived counterfactual projections of the 2023 resurgence extrapolated from pre-pandemic epidemic cycles. Under non-pharmaceutical interventions (NPIs), MP disappeared more completely and for longer than influenza or RSV; it returned along a policy-timed interval gradient — 6–7 months after NPI relaxation in China, 21–22 months in Japan — exceeding the no-COVID counterfactual projection by 19.2%, breaching its upper prediction interval. Modeled MP mortality in children aged 0–19 years (26,382 deaths, 95% uncertainty interval (UI) 18,984–36,012; ninth of 26 modeled lower respiratory infection (LRI) etiologies) concentrated where measurement is absent: 81.6% of estimated deaths in sub-Saharan Africa and South Asia, while infections peak at school ages where death is rarest — an infection–mortality dissociation that no single instrument of global health accounting can see. Chronic sequelae compound the omission: bronchiolitis obliterans developed in 7.3% of the largest post-pneumonia follow-up cohort, converting a survivable infection into decades of disability invisible to mortality-weighted metrics. The 2023–2024 resurgence was not an anomaly but an audit finding — a structural failure of measurement, not of biology. We propose five corrections with named institutional owners: MP-specific sentinel surveillance, sequelae-weighted burden metrics, resistance-stewardship infrastructure, diagnostic adoption standards, and vaccine prioritization.