Twenty-one cynomolgus macaques exposed to Bundibugyo virus died at a rate of 67% in a bioRxiv preprint posted on 10 August, the first natural-history model published for the species1. A natural-history model is an animal study that maps how a disease progresses: what it kills, how fast, and what it leaves behind. Disease in these animals ran a more protracted course than the published courses for Zaire and Sudan ebolavirus. The virus targeted macrophages and dendritic cells early. Animals that survived cleared systemic inflammation by day 28 but retained viral antigen in immune-privileged tissue, sites such as the eye and testicle where the immune system does not fully reach, with inflammation alongside it.
That 67% is a property of the model and of nothing else. The macaques were dosed to a schedule chosen to make the study informative for countermeasure work, at an exposure no person receives. In DR Congo, 46.8% of confirmed human patients have died, a figure measured on people who reached a laboratory rather than on animals given a fixed dose. The two numbers describe different systems and do not belong on the same axis. Neither one predicts the other, and the macaque figure says nothing about a Congolese patient's odds.
Regulators read a vaccine or drug trial against a characterised animal model. Without an agreed lethality baseline and a known disease course, an experimental countermeasure has nothing to be measured against, which is why the model comes before the trial rather than after it. The randomised treatment trial that dosed its first patient in Ituri on 2 July did so without one. For a species with three recorded outbreaks in nineteen years, nobody had built the model until now, and it arrived in the outbreak's fourth month.
