CIDRAP reported that a child in Bangladesh died on 1 February 2026 after household poultry contact, infected with a viral clade that has circulated in local poultry since around 2011 . Cambodia has logged three H5N1 human cases through 2026 to date. The global 2026 H5N1 human-case total outside the US dairy pathway sits at four. None of these involves dairy cattle, and all involve traditional backyard or smallholder poultry exposure.
South and Southeast Asian poultry H5N1 has been an endemic, household-scale threat for more than a decade. The clades involved tend to descend from Goose/Guangdong-lineage viruses, with regional reassortants in Bangladesh and Cambodia that diverged from the European and American clade 2.3.4.4b lines years ago. Exposure typically follows a recognisable pattern: a child or carer slaughters, plucks or feeds infected birds in a domestic compound, and the virus enters via mucosal contact rather than aerosol. Mortality among confirmed human cases on this pathway runs above 50% historically, far higher than the US dairy cohort, where workers have largely presented with conjunctivitis and mild respiratory symptoms.
The two threads matter for surveillance arithmetic. A bare global tally of H5N1 human cases lumps the Americas dairy cohort with the persistent Asian poultry pipeline and presents the trend as one curve. The structural picture is two curves on different y-axes: a wider mammalian dairy reservoir on one side, with low individual lethality but novel host range, and a long-running bird-to-human poultry pathway on the other, with higher individual lethality but a well-understood transmission ecology. Each pathway calls for a different intervention design. Smallholder poultry biosecurity in rural Bangladesh and Cambodia is a development and veterinary-extension question; California milking-parlour ventilation is an occupational-health and PPE question.
WHO and national agencies write monthly H5N1 risk assessments off cumulative case counts; if those counts are presented without the structural split, the global signal can read as runaway when one half is structurally novel and the other is a long-running endemic baseline. The separation also matters for vaccine prioritisation, where a stockpiled candidate matched to clade 2.3.4.4b will not necessarily protect against the Bangladesh or Cambodian lineages. Calibrated reporting, in this case, is not a stylistic preference but a precondition for sensible policy.
