Policy – July 29, 2026
ISA illustration: SalmonBusiness
Most of the mathematical models used to understand how infectious salmon anaemia spreads between farms have not been properly tested against real-world outbreaks, according to a major review of the available evidence.
The review nevertheless found a consistent pattern across existing research: infection risk falls as the distance travelled by seawater between salmon farms increases. But connections created by shared boats, equipment and personnel can limit the protection gained by moving sites farther apart.
The findings are reported in A Systematic Review of Studies Investigating the Transmission Dynamics of Infectious Salmon Anemia Virus, published in Reviews in Aquaculture on July 21.
The paper was written by Ahsan Raquib, Persia Carol Thapa, K. Larry Hammell, Javier Sanchez, Nicole O’Brien and Krishna Kumar Thakur.
The researchers identified 898 records through four databases and Google Scholar, before adding another five studies found through manual searches. After removing duplicates and screening the results, 24 modelling studies were included in the review.
Infectious salmon anaemia, or ISA, is a potentially devastating viral disease of Atlantic salmon. Severe outbreaks can kill more than 90% of fish at an affected site. The Chilean epidemic beginning in 2007 caused losses estimated at around $2 billion.
Across the studies examined, the risk of infection consistently fell as the seaway distance between farms increased.
One Norwegian study estimated that, compared with a farm immediately beside an infected site, relative risk fell to 20% at three kilometres, 13% at five kilometres, 6% at 10 kilometres and 2% at 20 kilometres.
These figures measure the distance travelled by water rather than a straight line on a map. Islands, coastlines, tides and currents can therefore determine whether two farms are genuinely connected.
However, the review found that belonging to the same local contact network also increased the risk of infection. Such networks can be created when farms operated by the same company share boats, equipment or personnel.
The authors said separating two farms by more than five kilometres may produce only a limited reduction in risk if they remain connected through these operations.
They also found that more than 30% of cases in some studies could not be linked to a documented source, suggesting that important routes of infection remain poorly understood.
The reliability of the models themselves was a central concern. Only four of the 24 studies reported validating their models, mostly by asking experts whether their structure and results appeared credible. Independent testing against real-world disease behaviour remained uncommon.
Most models also failed to include several possible transmission routes, including movements of live fish, wellboats, shared equipment and staff working at multiple sites.
The authors said future models should make greater use of real outbreak data and account for seasonal currents, wind, viral strains and operational links between farms.
They concluded that rapid culling of infected sites remains the most effective way to prevent ISA spreading to neighbouring farms.