Infectious salmon anemia in vaccinated Atlantic salmon: a field case study

1 Introduction

Infectious salmon anemia (ISA) was first detected in Norway in 1984 (Thorud and Djupvik, 1988). The disease was later detected in all major countries producing Atlantic salmon (Salmo salar L.) (Bouchard et al., 1999, 2001; Godoy et al., 2008). The number of annual outbreaks in Norway has varied from 80 at its peak in 1990 to 10–25 outbreaks annually over the last 10 years (Moldal et al., 2026). The introduction of general and specific biosecurity measures in 1989–1991 led to a rapid reduction in the annual number of cases.

The causative agent is infectious salmon anemia virus (ISAV), belonging to the species Isavirus salaris within the genus Isavirus in the family Orthomyxoviridae (Dannevig et al., 1995). ISAV occurs in two basic genotypes, North American and European (Blake et al., 1999; Kibenge et al., 2001). The European genotype is more widespread, showing higher genetic and virulence variation than the North American genotype (Krossøy et al., 2001; Kibenge et al., 2009; Plarre et al., 2012). Furthermore, ISAV has two main phenotypic variants based on the sequence of the highly polymorphic region (HPR) of gene segment 6: non-pathogenic ISAV HPR0 and pathogenic ISAV HPRΔ (Cook-Versloot et al., 2004).

It has been demonstrated that primary cases of ISA occur when ISAV HPR0 transitions to ISAV HPRΔ through a deletion in HPR and a mutation or insertion in the cleavage site of gene segment 5 (Fourrier et al., 2015; Christiansen et al., 2017; Cardenas et al., 2020). ISAV HPR0 is highly prevalent in Norwegian farmed salmon and occurs in hatcheries, marine on-growing farms, and brood stock facilities (Lyngstad et al., 2012; Moldal et al., 2025). ISA has only been detected in farmed Atlantic salmon, while ISAV has been detected in wild salmonids and farmed rainbow trout (Oncorhynchus mykiss) under natural conditions (Nylund et al., 2019; Alarcon et al., 2021).

ISA may cause high mortality and significant economic losses. Cumulative mortalities in Atlantic salmon during natural ISA outbreaks and experimental infections range from 0 to 100%. Typical clinical signs in infected fish are ascites, congestion and enlargement of the liver and spleen, congestion of the intestine, and petechiae in the peritoneum (WOAH, 2022). The hematocrit values may drop to less than 10 toward the end of a protracted disease course. Infection with any form of ISAV is listed by the World Organization for Animal Health (WOAH, 2022), while only infection with ISAV HPRΔ is listed within the European Union (EU, 2018).

A primary outbreak may cause secondary outbreaks due to horizontal transmission of virus to nearby farms through live animal movements, vessel movements, or passive transport via water currents (Murray et al., 2002; Lyngstad et al., 2008). Due to its status as a listed disease, ISA outbreaks in Norway are managed with mandatory disease control measures including the establishment of a restricted zone with surveillance of fish populations and usually culling or harvesting of the entire population on the affected site (NFSA, 2025). The regulatory authorities in Norway aim to decrease the number of annual outbreaks with stringent management, resulting in the culling of fish upon confirmation of ISA. In 2025, the Norwegian Food Safety Authority (NFSA) changed ISA management toward a risk-based, cage-by-cage approach, allowing continued production in ISAV-free cages based on a thorough monitoring and screening program (NFSA, 2025).

Routine husbandry stressors such as mandatory treatments for sea lice may play a critical role in ISA disease dynamics in a population. The impact of sea lice in Norway is regulated by implementing a legislative threshold of lice levels that dictate when delicing treatments must occur to keep the louse level below certain thresholds (Fiskeridepartementet, 2009). This level is an average of 0.2 adult female lice per fish during the wild salmon migration period in the spring and 0.5 adult female lice per fish during the rest of the year. As a result of this regulation, the majority of farmed Atlantic salmon are subject to several delicing treatments per year, which may consist of one or more delicing methods, including cleaner fish and chemical, thermal (in this article termed Thermolicer), or mechanical treatments. However, the latter two are associated with increased stress, injuries, and mortality that may result from handling and crowding during treatment (Overton et al., 2019; Ringstad et al., 2025). This may further worsen a situation with a latent infection in a population, resulting in an increased number of diseased individuals, pathogen shedding, and mortality.

Biosecurity measures to prevent introduction and spread of pathogens include screening of roe, separation of generations in space and time, routine fallowing, and vaccination (Lillehaug et al., 2015). Vaccination has been crucial to the success of fish farming in Norway and on a global scale. Although vaccination against bacterial diseases has been highly effective, this approach has been more challenging for viral diseases because viruses are intracellular.

For control of ISAV, in addition to the general biosecurity measures, it has been shown that using fish bred for increased resistance to ISA combined with vaccination against ISAV may reduce the probability of infection, but transmission was not prevented (Chase-Topping et al., 2021). In Norway, ALPHA JECT micro 7, a vaccine containing ISAV HPRΔ in combination with infectious pancreatic necrosis (IPN) virus and several bacterial antigens, is available (PHARMAQ, 2024). According to the leaflet for the vaccine, it offers immunity within 500 degree-days, lasting for at least 7 months postvaccination, and vaccinated fish would display reduced or no signs of clinical disease. However, they can still be infected, shed virus, spread the virus to other fish, and become asymptomatic carriers (Chase-Topping et al., 2021). Several outbreaks in Norway in recent years have been detected in vaccinated populations (Moldal et al., 2026). Different countries have adopted different strategies for ISA vaccination, such as mandatory vaccination of the whole population, voluntary regional control measures, or prohibition of vaccination. Vaccination against ISAV is mandatory in Chile and voluntary in Norway. The USA, with a history of ISA, relies on a surveillance and control program for regulation of the disease, with eradication as the official goal (APHIS, 2026). While not stated as an explicit ban, ISA vaccination is not permitted in the UK, which has ISA-free status (Scottish Government, S, 2023; CEFAS, 2024).

ISAV detection, resulting in outbreak affecting four marine farms, was reported in 2020 in southern Newfoundland, Canada. Atlantic salmon were vaccinated against ISAV, and the study reported the first description of epidemiologic characteristics of an ISAV outbreak with several ISAV variants at the population level (Romero et al., 2022). It is not clear whether the fish populations at the Canadian farms had completed the required immunization period before seawater transfer.

In Norway, a population of Atlantic salmon that had completed the required immunization period before seawater transfer was diagnosed with an ISA outbreak within a few weeks after seawater transfer in June 2021, with minimal clinical signs and no specific increase in mortality. To build further knowledge about ISAV infection dynamics in a vaccinated population under field conditions, following dispensation from the Norwegian Food Safety Authority to continue production at the site under close monitoring, we performed systematic surveillance using several sample matrices. Here, we describe infection dynamics, viral shedding, and clinical manifestations and evaluate environmental nucleic acid (eNA) analysis using swabs and water samples during a full production cycle in seawater.

2 Materials and methods

2.1 Fish

The fish of the SalmoBreed strain originated from a hatchery owned by Salaks Holding AS, a family company that also owns the sea site Salangslia in Troms County, Northern Norway, where this study was carried out. The fish had been vaccinated with ALPHA JECT micro 7 containing the European ISAV genotype approximately 20 weeks before sea transfer (range 1,573–1,926 degree-days). A total of 754,825 Atlantic salmon (Salmo salar L.) weighing approximately 270 g were transferred to seawater from 6 to 17 May 2021 and distributed among seven cages (Table 1; Figure 1) for grow-out at Salangslia.

Table 1. Average weight of salmon smolts at the time of seawater transfer for each cage.

The average weight is based on the operational weight at the cage level during seawater transfer and not on the weights of individual fish.

Of the 1 million smolts designated for seawater transfer, approximately 150,000 smolts of inferior quality were sorted using a fish-grading grid with a 17.5 mm opening and culled at the hatchery; thus, only 85% of the original fish were considered robust and well developed for seawater transfer.

The seawater site was situated in a protection zone within an ISA restriction zoneanemia due to an earlier outbreak of ISA in the area (Figure 1). According to the ISA control measures implemented by the NFSA, monthly sampling of heart and kidney tissues from 10 fish for ISAV screening using real-time reverse transcription PCR (RT-PCR) is performed at all farms within a restriction zone to control the spread of ISAV HPRΔ out of the zone. Samples taken by the fish health company Åkerblå, now a DNV company, for routine screening at the start of June 2021 tested positive for ISAV HPRΔ by PCR, and thus, on 16 June 16, the status of the location was changed to suspicion of ISA. Inspection and new sampling by the NFSA and analysis at the Norwegian Veterinary Institute (NVI) confirmed ISA on 28 June 2021. In accordance with the regulations in force at the time, the NFSA would decide that the fish at the site should be culled within a given time frame, normally within 2 months, with ISAV-positive cages removed from the site first.

The accumulated mortality at the site from seawater transfer to the suspicion of ISA in June 2021 was 0.35%. During the period from the detection of ISA in June until the end of August 2021, while awaiting a final decision from the local NFSA on the deadline for site clearance, Salaks Holding AS performed active surveillance. Samples were collected from 70 moribund fish and 575 apparently healthy fish and analyzed for ISAV using real-time RT-PCR at Pharmaq Analytiq AS. Only three moribund fish, one from C2 and two from C3, tested positive for ISAV. As the disease situation appeared to be stable and the fish had completed the recommended immunization period before seawater transfer, the company decided to apply for continued production at the site and offered to undergo active surveillance. Dispensation from the ISA regulation was granted by the NFSA based on a plan for close follow-up with defined criteria for termination (Supplementary File 1). A field study approved by the NFSA was initiated in September 2021. The timeline of the major events before the start of the study and until the slaughter of the fish is shown in Figure 2.

2.2 Dispensation to hold the fish at the site

An outbreak of ISA at a site normally leads to mandatory slaughter or culling of fish in affected cages within weeks and in all remaining cages within a few months. In this case, the NFSA granted a dispensation based on scientific advice from NVI and the Institute of Marine Research, Norway (NFSA, Dispensation permit 2021/133949, issued on 22 September 2021). Several conditions were predefined to prevent high viral shedding from the fish population, which would represent a risk to neighboring farms. These included an extensive monitoring and sampling program, assessment of mortality development, necropsies of moribund or dead fish, PCR-based ISAV screening, and monitoring of hematocrit in randomly collected fish from all cages. Termination criteria were agreed on based on the monitoring and screening results, triggering immediate harvest or culling of the affected cage if the given threshold was exceeded.

2.3 Fish health inspection and clinical observations

Fish health experts from Åkerblå visited the farm twice a month throughout the study period to perform risk evaluations, clinical inspections and sampling and issued a status report after each visit. These visits were often combined with the monthly sampling described later. Sampling during the physical inspections was risk-based, focusing on the cages with the highest probability of finding ISAV-positive fish. However, all cages were inspected at every visit, with deviations noted and moribund fish and fish displaying abnormal behavior or other clinical signs of disease sampled for necropsy. During these routine inspections, additional fish showing signs similar to those of ISA, such as protruding eyes, cutaneous bleeding, bleeding in the eye, petechiae, and ascitic fluid in the peritoneum, were sampled. Clinical signs and pathologies that were attributed to other diseases were also accurately described in the reports.

2.4 Sampling and handling of samples

2.4.1 Fish

A total of 2,821 fish, of which 1,690 were apparently healthy and 1,131 were moribund or dead, were sampled at the site during this case study (Table 2). Every 4 weeks from September 2021 to October 2022, 16–20 apparently healthy fish from each cage were randomly collected and sampled. In addition, 80 moribund or dead fish were sampled during the period between the scheduled samplings. Fish sampled between the dedicated sampling points were chosen from among moribund and dead fish removed daily from the site and analyzed such that all cages were represented in the analysis. The sampling protocol for apparently healthy and dead/moribund fish is shown in Figure 3.

Table 2. Number of individual fish sampled per month at the sea site, including apparently healthy and moribund/dead fish.

Variation in the number of fish per month reflects the fact that sampling was performed every 4 weeks. For example, in March, there were two sampling events in addition to sampling of moribund fish during the intervening 4-week period.

The criteria for choosing moribund fish for sampling were as follows: (1) sampled fish should represent all cages at the site over the period between samplings every 4 weeks; (2) fish should be chosen randomly if more fish were available than required; and (3) if there were not enough moribund fish, fish that appeared to have died recently were sampled. Sampling of moribund/recently dead fish was performed continuously based on the daily situation to enable the earliest possible observation and detection of fish potentially affected by ISAV.

Swabs from the second gill arch and the skin below the dorsal fin were collected in individual tubes with NucliSENS easyMAG Lysis buffer (bioMérieux) and stored at 4°C if samples were to be transported within 24 h (Aamelfot et al., 2015). If transport within 24 hours was not possible, the swab samples were frozen at -20°C on site and transported frozen. Blood was withdrawn from the caudal vein into heparinized tubes for hematocrit measurements, which were successful for 1,505 of the apparently healthy individuals and 970 of the moribund fish. Tissues from the heart and mid-kidney were collected in individual tubes with RNAlater and stored at 4°C.

Sampling for histology and immunohistochemistry differed depending on the status of the fish being sampled. For healthy fish, gills, heart with atrium, mid-kidney, and liver were sampled, while gills, heart with atrium, skin with muscle, pyloric caeca with pancreas, liver, mid-kidney, spleen, and pseudobranch were sampled from moribund and dead fish. All samples were fixed in 10% buffered formalin. Additionally, heart and mid-kidney were collected in transport medium for the purpose of possible propagation of virus.

To avoid potential bias from processing and analysis at only one laboratory, samples from approximately 50% of individuals were analyzed at NVI, while the rest were analyzed at Pharmaq Analytiq, Norway. Anesthetized fish were randomly assigned to two parallel sampling stations and sampled simultaneously, with one station sampling for each laboratory.

In total, 20 randomly selected fish per cage among those sampled every 4 weeks were weighed by the farming company to monitor growth. The average weight of 20 fish per cage, representing 140–160 fish at each main sampling time point, was calculated.

A total of 840 fish were sampled at the slaughterhouse from July to November 2022. Sampling at the slaughterhouse was distributed such that 40 fish were sampled each time a proportion of a cage was slaughtered, bringing the total to 120 fish per cage. All samples from the slaughterhouse were analyzed at NVI. At the slaughterhouse, gill and heart tissue samples were collected in RNAlater and stored at 4°C. Swab samples were also collected from various surfaces at three points: (1) where the fish underwent percussion and a cardiac cut for killing; (2) the exsanguination tank; and (3) the tray where fish were transported after exsanguination while awaiting further processing.

All samples were transported to the laboratory within 24–72 h. Samples for PCR analysis were processed immediately or frozen at -20°C upon arrival.

2.4.2 Water samples and filtration

Water samples were collected monthly at the water surface from each cage (Figure 4, top panel), during delicing (Figure 4, bottom panel, and Figure 5), and from well boats after transport for harvest.

During delicing, water samples were collected at four points (Figure 4, bottom panel): (1) in the cage after the fish were first contained; (2) in the well boat after the fish had been pumped aboard and allowed to settle for 2–2.5 h, before being pumped through the Thermolicer; (3) where the fish were pumped out of the treatment chamber; and (4) in the cage after the fish were pumped back into their respective cages.

At harvest, water samples were collected from the well boats immediately before the fish were pumped into the slaughterhouse. In Norway, when fish are transported from a farm with ISA for slaughter, they should be transported and held in a closed transport system. Fish from this farm were slaughtered over a period from June to October 2022. Each time fish were pumped into a closed well boat, transported to the slaughterhouse, and held in the well boat for several hours until slaughter, water samples were collected immediately before the fish were pumped into the slaughterhouse.

Then, 1 L of water was sampled in a sterile flask and kept cool with a freezing element. Water samples were shipped the same day by overnight transport and filtered immediately upon arrival at the laboratory.

The sea site, Salangslia, has a green aquaculture license, meaning that the level of adult female sea lice per cage must be below 0.25 throughout production. This resulted in several delicing operations for this population, with seven thermal treatments during the study period. Since stress due to handling and delicing treatments might reduce the general immune status of salmon, enabling infection to flourish, we collected water samples during six of the delicing operations to test for possible viral shedding. However, at two of the time points, the samples were delayed in transport for more than 4 days and were therefore considered unsuitable for analysis. Figure 5 gives an overview of the timeline and types of delicing treatments at the farm site.

At the laboratory, water samples were processed for concentration and analyzed using a method published earlier by Weli et al. (2021). Briefly, 1 L of each water sample was filtered using a 47 mm electronegatively charged membrane filter (MF-Millipore 0.45 μm MCE membrane, Merck Millipore) placed in a 47 mm in-line filter holder (Merck Millipore) with a peristaltic pump (V6-3L Peristaltic Pump, Shenchen) at a flow rate of 200 mL min⁻¹. After filtration, the filter was placed in a 50 mm Petri dish and eluted with 2 mL of NucliSENS Lysis Buffer. The eluted buffer was divided equally between two tubes and processed for analysis. For water samples from delicing treatments, more than one filter was often necessary due to substantial amounts of particles in the samples. In this case, all filters were combined during elution. The eluted samples were immediately frozen at -20°C until further analysis by PCR.

2.4.3 Extraction of nucleic acids, PCR, and sequencing

At NVI, tissues from heart and gill were homogenized individually in MagNA Pure RNA Tissue Lysis Buffer (Roche) at 25 Hz for 10 min, and 50 µL of the lysate was transferred to a MagNA Pure 96 Processing Cartridge (Roche), followed by the addition of 450 µL phosphate-buffered saline (PBS). For swabs and water samples, 500 µL of material (namely, lysis buffer) was transferred directly to a MagNA Pure 96 Processing Cartridge (Roche). Nucleic acids were extracted with a MagNA Pure 96 Instrument (Roche) using the MagNA Pure 96 DNA and viral NA Kit (Roche) and the Pathogen Universal 500 protocol. The concentration and quality of RNA were measured using a NanoDrop ND-1000 UV-Vis Spectrophotometer (NanoDrop Technologies), and 5 µL of template was used for real-time RT-PCR detection of ISAV with primers targeting segment 8 (Snow et al., 2006). Samples with a Ct value below 40 were amplified with primers for ISAV HPR (WOAH, 2022). PCR products of the expected size were sequenced with the same primers.

At Pharmaq Analytiq, RNA was extracted using a custom version of the Omega Bio-tek Mag-Bind^® Total RNA 96 Kit according to the manufacturer’s recommendations. Extracted total RNA was eluted in a final volume of 100 µL of the supplied kit elution buffer. Extracted RNA from salmon tissues and swabs was tested by TaqMan real-time RT-PCR with qScript XLT 1-Step RT-qPCR ToughMix (QUANTAbio). A housekeeping gene, elongation factor 1 alpha (EF1A), was used as an internal control (Olsvik et al., 2005), and a specific assay targeting segment 8 was used to detect ISAV (Snow et al., 2006). The specific assay was run in duplex reactions with EF1A. The primer and probe concentrations had been optimized and found to be 900 nM for primers and 225 nM for the corresponding probe for EF1A in all duplex assays. The primer and probe concentrations for ISAV were 600 nM for primers and 200 nM for the corresponding probe. The samples were run individually in standard 384-well plates in a total volume of 10 μL per well, with 2.5 μL of isolated total RNA as the template. Plates were analyzed in a QuantStudio 5 real-time machine under standard conditions. Each run consisted of 40 cycles, and the samples were considered positive when the fluorescence signal increased above a set threshold of 0.05. Samples that were PCR-positive for ISAV at Pharmaq Analytiq were sent to NVI for amplification and sequencing of HPR as described above.

2.4.4 Histology and immunohistochemistry

Formalin-fixed tissue samples were processed for histology and embedded in paraffin. Tissue sections (2 µm) were stained with hematoxylin and eosin (H&E), scanned with a Hamamatsu NanoZoomer S360 (NVI) or an Aperio ScanScope AT Turbo slide scanner (Pharmaq Analytiq), and evaluated digitally.

Immunohistochemical analysis for ISAV was performed on formalin-fixed tissue samples from individuals with positive PCR detection of ISAV in heart tissue using a polyclonal antiserum raised against ISAV nucleoprotein (anti-ISAV, K716) (WOAH, 2022). Immunohistochemistry slides were digitized and evaluated.

2.4.5 Virus propagation

Tissues in transport medium (minimum essential medium (MEM) with 10% fetal calf serum and 0.02% penicillin-streptomycin-amphotericin) from selected fish with ISAV detected in heart tissue by PCR were processed and inoculated into Atlantic salmon kidney (ASK) cells (WOAH, 2022). Virus identification was performed using Indirect Fluorescent Antibody Test (IFAT) with the anti-ISAV monoclonal antibody P10 (Aquatic Diagnostics Ltd., Stirling, Scotland).

2.5 Data analysis and graphical presentation

To identify possible macroscopic findings during autopsy that might lead to optimized selection of fish for ISAV detection, we explored associations between hematocrit, pathology, and Ct values.Hematocrit (HCT) was measured successfully in 2,475 of the 2,821 fish sampled. The 2,475 fish were grouped as follows: 155 apparently healthy fish with no autopsy findings (A), 1,350 apparently healthy fish with autopsy findings (B), 18 moribund fish with no autopsy findings (C), and 952 moribund fish with autopsy findings (D). The HCT values in the apparently healthy fish without autopsy findings were used to determine a normal range of HCT and set 95% ( ± 1.96 × SD) boundaries for anemia (HCT below 31) and hemoconcentration (HCT above 59), respectively. Anemia was further classified as very severe (HCT below 10), severe (HCT 10–20), or moderate (HCT 20–30).In group D (moribund fish with autopsy findings), odds ratio (OR) was used to calculate the association between ISAV HPRΔ detection in the heart and in all fish samples (heart and swabs) combined and the presence or absence of anemia, skin ulcers, pale organs, internal bleeding, dark liver, ascites, swollen spleen, and hydropericardium.

The Ct values for all samples were entered into GraphPad Prism version 9 for graphical presentation. A total of 128 fish were included in the analysis to calculate the detection proportion with 95% confidence intervals using the Wilson method.

Differences between sample matrices (heart, gill swab, and skin swab) were assessed using Cochran’s Q test for paired binary data, followed by pairwise McNemar’s tests. Comparisons between moribund and apparently healthy fish within each matrix were also conducted using McNemar’s tests, with a significance level of p < 0.05.

3 Results

3.1 General fish health

As described previously, the main focus of the fish health services was to observe clinical signs of ISA throughout the production period. During the 36 visits by the fish health services and their participation in the regular sampling, the fish health experts did not observe clinical signs of ISA in the autopsied fish. In some of the reports, a few individuals were reported to have dark-colored livers. These individuals were sampled in addition to the planned monthly samples of apparently healthy and moribund fish, but ISAV was not detected in these fish.

The fish weighed approximately 1,300 g in September 2021 and increased in weight over the production period (Figure 6). However, from January to the end of March 2022, the growth of the fish in all cages was slow. From June 2022 until slaughter, growth accelerated.

The accumulated mortality from seawater transfer in May to December 2021 was 1.9%. In this period, the fish had undergone three Thermolicer treatments. The monthly mortality increased to approximately 0.8–1% from February to April 2022, as shown in Figure 7. Mortality was 2.2% in May and 1.2% in June, decreased to 0.44% in July, and increased to 0.7–0.9% in August and September before the fish were slaughtered in October. Total accumulated mortality at this site during the production cycle reached 10.04%, which can mainly be attributed to general production-related factors, including handling, delicing, and the presence of multiple bacterial, viral, and parasitic infections, while no direct association with ISA was identified.

The number of fish with abnormalities increased, particularly after salmon lice treatments during winter 2022, with affected fish often exhibiting skin ulcers. Other abnormalities and signs of other diseases were also described in the reports.

In addition to ISA, heart and skeletal muscle inflammation (HSMI), parvicapsulosis, epitheliocystis, and winter ulcers (Moritella viscosa) were diagnosed, resulting in a complex disease situation at the site (Table 3). Ulcers were observed from mid-December, and causative agents were confirmed as Moritella viscosa and Tenacibaculum sp. Ulcers and sepsis were observed in three of the cages in January 2022. By March, fish with ulcers, HSMI, and epitheliocystis were confirmed in all cages.

Table 3. Disease status (other than detection of ISAV) at the farm during study period.

No diseases were reported during inspections until November 2021, and in September 2022, no specific diagnosis was made for the fish that were inspected.

Later in March 2022, fish in two of the cages, C5 and C10, were affected by parvicapsulosis. The situation within the population continued to become more complex, and in June 2022, several fish from cages C4, C9, and C11 were severely affected by costiasis, and a few fish in C3 also showed signs of nephrocalcinosis. Toward the end of the production cycle, the frequency of fish with ulcers appeared to decrease, while fish continued to show clinical signs consistent with HSMI, costiasis, and epitheliocystis.

3.2 Analysis of samples

3.2.1 Tissue and swab samples

In total, 3,661 fish were sampled during the case study, with 2,821 fish sampled at the sea site and 840 fish at the slaughterhouse. The monthly distribution of positive detections is shown in Figure 8. The number of individuals with ISAV detected using PCR increased from April 2022. Of the 2,821 fish sampled at the sea site, 128 tested positive for ISAV in at least one of the tissue or swab samples. Of these 128 individuals that were ISAV-positive using real-time RT-PCR, at least one of the sample matrices from 60 individuals were sequenced, while it was not possible to amplify the rest of the samples. One of these 60 samples was confirmed to be ISAV HPR0 (in November 2021), and the rest were ISAV HPRΔ. At the time of ISA confirmation in June 2021, positive individuals were detected in four (C2, C3, C4, and C12) of the seven cages, while toward the end of the production cycle and at the time of slaughter, ISAV HPRΔ was detected in fish from all seven cages.

The Ct values for each sample type from the individuals taken at the sea site are plotted in Figure 9. The number of individuals testing positive with skin and gill swabs was higher than that with heart tissue samples (Table 4). However, when the samples from the same individuals were positive, the difference in Ct value was not significant between the heart tissue and swabs (data not shown).

Table 4. Number of ISAV-positive individuals in different sample matrices taken at the sea site during the study period.

Only a few fish tested positive through March 2022, of which only one, sampled from cage 3 in December 2021, was sequenced and confirmed to be ISAV HPRΔ. In April 2022, the number of individuals with ISAV detection increased, with the majority of detections occurring in the cages that had detections at the start of the study and were located nearest to the shore (Figures 1, 10). The number of individuals testing positive for ISAV after April 2022 increased in the majority of cages, with a slight decrease in May 2022, until the fish were slaughtered. Among the ISAV-positive samples, the number successfully sequenced and characterized as ISAV HPRΔ increased after April 2022. Altogether, five fish were diagnosed with ISA, all between June and October 2022, toward the end of the production cycle. These fish were sampled from cages 2, 3, and 11.

ISAV HPR0 was detected in one gill swab in November 2021. None of the other samples that were positive for ISAV in November 2021 or in February and March 2022 could be sequenced because template amplification was insufficient. ISAV HPRΔ was detected in heart tissue and a gill swab from one fish sampled in early December and in several samples from several fish sampled from April onward until the sea site was emptied by the end of October 2022. A higher number of individuals tested positive for ISAV in skin swabs, followed by gill swabs and then heart tissue (Table 5). Only 18 fish were positive in all three sample matrices. Detection proportions differed significantly between sample matrices (Cochran’s Q = 52.38, df = 2, p < 0.0001). Skin swabs exhibited the highest detection (73%, 95% CI 65.2–80.2), followed by gill swabs (41%, 95% CI 33.2–50.0) and heart samples (25%, 95% CI 18.2–33.2) (Table 6). Pairwise McNemar’s tests indicated significant differences between all matrices (Table 7). Discordant pair analysis showed that gill swabs detected more positive cases than heart samples, while skin swabs detected substantially more positives than both gill swabs and heart samples. Detection probability was consistently higher in moribund than apparently healthy fish across all matrices (p ≤ 0.001), indicating a strong effect of disease status. At the slaughterhouse, ISAV was detected in at least one of the samples from 77 of the 840 fish sampled (Figure 11). When comparing the ISAV detection between cages, it was clear that cage 2, which was also slaughtered last, had the highest number of positive detections. Sequencing HPR from slaughterhouse samples demonstrated ISAV HPRΔ in samples from 17 fish (Table 5). Gills, swabs from skin and gills, and heart tissue were sampled at the slaughterhouse, and ISAV was detected more frequently in gill tissue than in heart tissue. The swab samples from both gills and skin were more frequently positive for ISAV than the gill and heart tissue samples.

Table 5.

Symbol # - Number. NA - not applicable as gill tissues sample were not collected.

Table 6. Detection proportion by sample matrix.

The number of ISAV-positive fish (n) among the total of 128 positive fish and the corresponding detection proportion are presented. CI, confidence interval.

Table 7. Pairwise McNemar’s test comparing detection probability between sample matrices.

Test statistics (χ²) and corresponding p-values are reported for each comparison. Significant p-values (p < 0.05) indicate differences in detection probability between matrices. GS, gill swab; SS, skin swab.

At the slaughterhouse, environmental swab samples were also taken from three surfaces as described earlier. Among the samples taken at the time of slaughter of the last batch from cage 4 in September 2022, the swab sample from the surface of the exsanguination tank was ISAV HPRΔ-positive. This was the only ISAV-positive environmental surface swab sample among 12 samples taken over three time points for cage 4. For cage 2, which was the last cage that was harvested, 11 of 12 environmental surface swab samples distributed over four days were ISAV-positive. Of these 11 positive samples, 9 were confirmed as ISAV HPRΔ with an average Ct value of 30.9.

3.2.2 Water samples

A total of 98 water samples tested positive for ISAV among 364 samples collected at the farm site. An overview of positive samples over the study period is shown in Figure 12. Only 21 of these 98 ISAV-positive samples were sequenced, with 20 confirmed to be ISAV HPR0 and one to be ISAV HPRΔ. The majority of the samples that were confirmed as ISAV HPR0 were from December 2021 to February 2022 and were from the samples collected during delicing operation. ISAV HPRΔ was detected in a water sample from C2 at the last sampling time point when the cage was being emptied of fish.

At the slaughterhouse, a total of nine water samples were ISAV-positive. The ISAV-positive samples comprised one from C10, two from C11, and six from C2. Only two of these nine samples were sequenced, both from C2, and confirmed to be ISAV HPRΔ.

3.2.3 Histological findings

In Norway, confirmation of ISA is based on virus detection using a minimum of two different biological principles, namely, histopathology and immunohistochemistry compatible with ISA and detection of ISAV RNA using PCR followed by sequencing that demonstrates ISAV HPRΔ. These criteria were met by a total of five fish during the sampling period. Four individuals showed kidney lesions with multifocal to diffuse interstitial hemorrhage with necrosis of tubular epithelium in affected areas (Figure 13A). Six individuals showed lesions in the liver with multifocal to confluent, zonal hemorrhagic necrosis (Figure 13B). ISAV was detected in the endocardium of five of these six individuals using immunohistochemical staining (Figures 13C, D).

Several other disease problems were diagnosed at the site during the trial period, including HSMI, cardiomyopathy syndrome (CMS), complex gill disease, parasitic gill inflammation (Ichthyobodo sp.), bacterial gill inflammation, epitheliocystis, parvicapsulosis, bacterial ulcers, and systemic bacterial infection (Table 3).

Epitheliocystis and HSMI were detected from March 2022, with a lower frequency of HSMI during the last months before harvesting. From May 2022, moderate to large numbers of Ichthyobodo sp. were detected on the gills, and pathological changes, including hyperplasia, inflammation, and vascular lesions, increased. CMS was detected in the latter part of the production cycle, from August 2022 until harvest. Parvicapsula pseudobranchicola was detected from the start of sampling until July 2022. Ulcers related to Tenacibaculum sp. and Moritella viscosa were observed from October 2021 throughout the production period, with a hiatus during summer 2022. In addition, systemic bacterial disease was diagnosed in a few individuals.

3.2.4 Association of hematocrit values and pathology findings with ISAV detection

An overview of the HCT values and their distribution among the sampled fish is given in Table 8. In group B (apparently healthy, with autopsy findings), there was considerably less anemia, 8.2%, compared with 62.1% in group D (moribund, with autopsy findings), which comprised the majority of the anemic fish and almost all fish with severe anemia. Group C (moribund without autopsy findings) was very small, with 11.1% anemia. Anemia was much more prevalent in moribund fish than in the other groups, with an observed plateau of more than 70% from January to June 2022 in moribund fish.

Table 8. Descriptive hematocrit (HCT) distribution in subgroups differentiated by clinical and autopsy status.

Healthy refers to apparently healthy fish before the autopsy was performed.

Fish in group D were further subcategorized according to the presence or absence of anemia, skin ulcers, pale organs, internal bleeding, dark liver, ascites, swollen spleen, and hydropericardium, combined with the presence of ISAV HPRΔ based on autopsy findings (Table 9).

Table 9. Odds ratio (OR) showing the strength of the association between descriptive hematocrit (HCT) and autopsy signs of ISA and ISAV HPRΔ detection in the 952 fish categorized as moribund upon sampling and in which hematocrit was measured.

confidence interval (CI), heart and swabs (H & S). Bold numbers represent statistical significance (p<0.05).

Single lesions such as skin ulcers, pale organs, and dark liver were not significantly associated with ISAV HPRΔ detection. Swollen spleen was associated with increased odds of detecting ISAV HPRΔ in heart tissue (OR 4.27, 95% CI 1.1–16.2, p = 0.032), but not when ISAV HPRΔ in swabs was included (OR 1.89, 95% CI 0.8–4.6, p = 0.162), indicating more pathology when the virus is detected in the heart.

Hydropericardium was found in only 13 of 952 fish with autopsy findings and was associated with high odds of detecting ISAV HPRΔ in heart tissue (OR 51.7, 95% CI 13.2–203.2, p < 0.0001).

3.2.5 Virus propagation

Virus was propagated from one of the 18 fish heart tissue samples in which propagation was attempted. The fish from which the virus was propagated was also confirmed to have ISA using PCR (Ct value in heart tissue, 15.74) and immunohistochemistry. This fish had a hematocrit value of 8, and autopsy revealed ulcers, an eroded dorsal fin, pale gills, adhesions, and melanization of internal organs, along with a bilobed and congested liver.

4 Discussion

An outbreak of ISA in vaccinated populations has been reported at four farms in Canada, with variation among the farms in the detection of ISAV over the production period (Romero et al., 2022). However, it is not clear whether the fish at each of these farms had reached full immunization before seawater transfer. To the best of our knowledge, this is the first paper describing detailed surveillance with several sample matrices from fish in an ISA outbreak. All fish in the farmed Atlantic salmon (Salmo salar) population were vaccinated and had attained full immunization. ISA was diagnosed shortly after sea transfer, and active surveillance was carried out over a period of more than a year from confirmation of ISA in June 2021 until the last fish was slaughtered in October 2022.

This study was conducted at a commercial farm and was initiated upon confirmation of ISA. Hence, inclusion of a control group of non-vaccinated fish in the cohort or a naïve fish group, which could have strengthened knowledge regarding virus transmission and infection of naïve fish, was not possible. The fish were not culled according to the regulation to build knowledge about the development of infection in vaccinated fish in the field. This study was carried out due to several factors, such as a low calculated risk of spread as there were few sites with fish in production in the area, and the nearest two sites with fish were 6–7.5 km away.

Fish were sorted at the smolt production site, where the fish that were considered robust were transferred to sea. Such a meticulous sorting process and the culling of 15% of the fish before seawater transfer are not common practices in salmon farming and may have influenced the results of this study. It can be speculated that the prevalence of ISAV-positive fish and mortality would have been higher if fish had not been sorted. Previously, a higher prevalence of ISAV-positive fish at four farms with ISA outbreaks in vaccinated fish was reported (Romero et al., 2022).

It is well known that handling fish undergoing infections will increase the risk of developing specific diseases and render them more susceptible to other pathogens and opportunistic microorganisms, including parasites. The fish health experts did not observe obvious indications of ISA in the autopsied fish during their inspections. However, since the fish were diagnosed with HSMI from March 2022, with pathological signs to some extent similar to those of ISA (ascites and blood in the abdomen), possible ISA signs could have been masked by HSMI clinical signs.

According to the fish health services responsible for several other locations in this region, the mortality observed in this study can be considered low compared with that at other similar farms in the production area. In December 2021 and at the end of January 2022, due to an increased salmon lice infestation, the entire fish group had to undergo two delicing treatments. Considered independently, this yielded positive results with relatively low acute mortality but resulted in ulcers in a substantial number of individuals, with Moritella viscosa detected in both ulcers and kidneys. The mortality was elevated at approximately 0.8–1% over the next 3 months (February–April). Yet another treatment in May, for the weakened fish that still struggled with winter ulcers, resulted in increased mortality to 2.2% postdelicing, probably due to low tolerance to handling. Two further treatments before the end of production in October led to an increase in mortality, altogether indicating that the impact of delicing on mortality was significant. The 4^th treatment in January probably resulted in extensive mechanical wounds and ulcers on the fish, which might have contributed to increased mortality with the 5^th treatment in May. The Salangslia site had a relatively high number of treatments against salmon lice: one chemical treatment with Slice in August 2021 before the study period started and seven Thermolicer treatments in a well boat. In comparison, according to the fish health services, the average number of treatments among 11 farms in the vicinity in 2021–2022 was 3.82. It has been reported that high lice burdens can be associated with multiple pathogenic bacterial genera (Vibrio, Flavobacterium, Tenacibaculum, Pseudomonas) in a salmon population (Llewellyn et al., 2017). Although several microorganisms capable of causing major diseases were detected, none of them alone seemed to have affected the total mortality.

Although handling stress due to delicing can be expected to influence disease development, it did not seem to affect the ISA disease situation at the site, as is often the case with viral diseases. The development of infections resulting in ulcers and gill infections caused by bacteria and parasites from January to June 2022, which caused a complex health situation at the site, may also have contributed to the lower growth of the fish during this period. Additionally, the increasing complexity resulting from the presence of infectious agents and diseases within a population might affect infection kinetics, disease development, and mortality. Co-infections with other agents and opportunistic parasitism may potentially affect the host’s immune system such that the prevalence of Desmozoon lepeophtherii may influence proliferative gill inflammation (Nylund et al., 2010). It has been reported that once the pathogens have become established in the population, they may persist for a very long time (Downes et al., 2018). However, these pathogens may persist and not necessarily have an impact on disease development and further mortality. This might be the case in our study, as we did not observe a sudden rise in mortality due to infections at any time, although several pathogens were established and morbidity due to HSMI was evident.

Tissues from the heart and kidney are recommended for detection of ISAV by PCR (WOAH, 2022). Virus was most prevalent in skin swabs, followed by gill swabs, with the lowest detection in heart tissue, suggesting that mucosal sampling may be a sensitive tool with a higher detection frequency. As swab samples likely reflect viral shedding through skin and mucosal surfaces from infected fish and pathogen circulation within the population rather than the infection status of individual fish, they may be particularly useful for early detection and non-lethal monitoring at a farm. Any detection of virus in such samples should be followed up by sampling the recommended tissues for the purpose of confirming disease according to established criteria.

ISAV was detected more frequently in samples from moribund fish than in samples from apparently healthy fish, an aspect that should be taken into consideration when choosing fish for sampling. Moribund fish were more likely to test positive for ISAV in heart tissue. Our results demonstrate a clear and consistent association between clinical morbidity, independent of causality, and the likelihood of detecting ISAV. Although such a relationship is expected given the pathophysiological effects of ISAV HPRΔ, the magnitude of the observed differences emphasizes the surveillance and diagnostic value of prioritizing clinically compromised fish over apparently healthy fish. These results imply that non-lethal sampling methods, particularly skin swabs, may substantially improve detection probability compared with heart tissue, with important implications for surveillance and early detection strategies.

Routine surveillance programs incorporating morbiditybased sampling strategies, especially with swab samples of moribund or very recently dead fish, can increase ISAV detection probability while reducing the unnecessary sacrifice of healthy fish. This approach may be especially valuable for early detection, when moribund fish can serve as sentinels of increasing viral activity within a population and support non-lethal sampling and the 3R (Replacement, Reduction and Refinement) principles.

After the initial detection of ISAV HPRΔ in summer 2021, it was detected in only one fish in December 2021 and then in several fish from April 2022 onward with a peak in July. It can be speculated that the effect of vaccination against ISA was less pronounced toward the end of the production period and/or that immunity in general was weakened due to complex health issues at the farm. Additionally, complex biotic and abiotic factors such as environmental conditions, stress due to handling, and multiple infections might influence vaccine efficacy, as has also been seen for piscirickettsiosis (Valenzuela-Aviles et al., 2022; Kumar et al., 2024). Altogether, only five fish were diagnosed with ISA according to the diagnostic criteria in Norway. The two methods used are real-time RT-PCR and immunohistochemistry. The latter has inherently low sensitivity and requires that the impacted niche in the tissue be sampled and sufficiently affected for detection. Thus, it can be argued that there might be a higher number of fish with ISA if diagnostic criteria were different and other more sensitive analytical methods were employed. In total, 32 of the fish sampled at the site were ISAV-positive in heart tissue using real-time RT-PCR, and of these, 17 could be sequenced and confirmed to be ISAV HPRΔ. In this study, the focus at the start of the study was on ISAV HPRΔ, and thus gill tissues for analysis of ISAV HPR0 were not included when planning the regular sampling at the farm. The results from the fish samples taken at the slaughterhouse show that gill tissues may be a good alternative, also for the detection of ISAV HPRΔ.

Although anemia traditionally serves as an indication of ISA in the fish, causes unrelated to ISA, such as ulcers and mechanical injuries, may also lead to anemia, as reflected in the hematocrit values in our field study. However, anemia may not be representative of ISA cases, which are often insidious. Thus, the use of anemia as a main indicator of ISA should be carefully assessed, especially in populations with other ongoing diseases.

Several common autopsy findings showed no or only borderline associations with ISAV HPRΔ detection. Some were biologically implausible and may reflect the complex disease situation in the population rather than a direct relationship with ISA. These findings should all be interpreted with caution as they originate from a complex field situation. The association of anemia (low HCT values) with lower odds of ISAV HPRΔ infection (OR = 0.36, 95% CI 0.20–0.66) illustrates this issue. This seems counterintuitive and may be explained by the association of low HCT values with several diseases other than ISA, resulting in an inverse association unlikely to represent a true protective effect. Although multivariable analyses are commonly used to account for confounding, the substantial dependencies among variables, the heterogeneous disease status of the population, and the complex, potentially non-causal relationships among clinical findings limited the potential for obtaining reliable and biologically meaningful estimates. Consequently, multivariable analyses were considered unlikely to provide additional insight and could potentially lead to misleading conclusions.

Hydropericardium stands out in this material as the most specific autopsy finding for ISA in moribund fish, with an estimated association with ISAV HPRΔ detection of OR = 51.3 (95% CI 13.2–203.2). This suggests that being observant for signs of hydropericardium during autopsy and prioritizing the sampling of these fish might increase the probability of ISAV detection. However, as no instructions were given to check specifically for hydropericardium, the 13 fish detected with this lesion may have had copious amounts of fluid in the heart sac, a known feature of clinical ISA. Perhaps excluding fish with only small amounts of fluid in the heart sac is necessary to obtain the high specificity of this lesion for ISA observed in this study. Further studies are needed.

Detection of environmental nucleic acids (eNA) in water samples has great potential as a surveillance tool for pathogens and other biological threats (Peters et al., 2018; Benedicenti et al., 2024). It has been demonstrated that salmonid alphavirus (SAV) can be detected using environmental RNA (eRNA) monitoring under field conditions at an earlier stage of the infection compared with traditional sampling methods (Bernhardt et al., 2021). In this study, only 20% of the ISAV-positive water samples collected from cages were successfully sequenced, and the majority of the samples were the ISAV HPR0 variant, while only one of 364 samples at the end of the study was characterized as ISAV HPRΔ. Despite being positive for ISAV using real-time RT-PCR, several samples with high Ct values could not be amplified for characterization as either ISAV HPR0 or ISAV HPRΔ, as also reported previously (Romero et al., 2022). It should be noted that pathogen detection in water samples is influenced by several methodological factors, particularly the volume of water used for concentration, pathogen load in the sample, and the limitations of the PCR assay. Thus, while pathogen detection may facilitate early intervention and management decisions, negative test results should not be interpreted as proof of pathogen absence and should be considered with caution in management contexts. For ISAV, simulations at a salmon farm site have shown that virus concentration varied spatiotemporally with the progression of the outbreak, current speed and direction, tidal elevation amplitude, and environmental decay (Ding et al., 2024). Moreover, other biophysical parameters such as temperature may affect the survival and/or infectivity of salmonid alphaviruses and ISAV, influencing the analysis of virus in water samples (Graham et al., 2007; Weli et al., 2025). The majority of samples in which virus was characterized were collected from well boats during delicing operations. This suggests that virus was either dispersed or reduced to non-detectable amounts while fish were kept in open sea cages, or that virus shedding from this population was low. Thus, water samples from semi-open cages appear to be less suitable for detection of ISAV HPRΔ from a population with low prevalence. ISAV surveillance using water samples in open and semi-open cages warrants further development, as detection sensitivity is strongly influenced by sampling, sample volume, and analytical workflow (Benedicenti et al., 2024). Water samples from thermal delicing machines have demonstrated Yersinia ruckeri infection in Atlantic salmon (Riborg et al., 2022). This is most likely due to a combination of a stress-related increase in shedding of the pathogen and the treatment of large numbers of fish in a limited volume of water, leading to the concentration of shed pathogens. This indicates some potential for ISAV monitoring in more enclosed systems.

5 Limitations of this study

This study was carried out during a normal production cycle at a sea site and thus lacked a non-vaccinated or naïve control group, preventing direct assessment of vaccine effects on ISAV transmission. In addition, the extensive pretransfer sorting and culling of approximately 15% of smolts, which are not common industry practices, may have reduced the observed prevalence of ISAV-positive fish and ISA-related mortality. Environmental nucleic acid (eNA) surveillance showed promise with swab samples for pathogen detection, but limited sequencing success and possible methodological constraints reduced the ability to characterize the detected virus, potentially leading to an underestimation of ISAV HPRΔ confirmation. Finally, the presence of multiple pathogens and co-infections may have influenced disease progression and mortality, making it difficult to attribute observed effects solely to ISAV.

6 Conclusion

Rigorous disease-specific surveillance combined with robust management strategies is critical for mitigating the risk of pathogen transmission to surrounding ecosystems and adjacent aquaculture sites, thereby reducing the potential for secondary outbreaks and preserving environmental integrity. Biosecurity measures such as reducing the introduction of ISAV HPR0 and ISAV HPRΔ and vaccination are crucial tools to reduce infections and disease outbreaks. The low prevalence of clinically confirmed ISA despite widespread viral detection is consistent with a potential mitigating effect of vaccination, although this study design does not permit direct assessment of vaccine efficacy. Several parameters may influence vaccine efficacy, particularly towards the end of the production cycle, when vaccine-induced immunity may have waned. An increase in the detection of ISAV-positive fish was observed toward the end of the production cycle. Active surveillance using swab samples, especially those from moribund fish, seems to be an important tool for the earliest possible detection of an ongoing infection in a population with low prevalence. Further understanding of optimal sampling practices combined with early detection will enable knowledge-based risk assessment and good management practices.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/faquc.2026.1885657/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Dennis B. Bela-ong, Gyeongsang National University, Republic of Korea

Reviewed by: Marco Rozas-Serri, Pathovet Laboratory, Chile

F. C. Thomas Allnutt, Independent Researcher, Glenelg, MD, United States

Mario Caruffo, University of Chile, Chile

Sunita Awate, UVAXX PTE LTD, Singapore

Magnus Vikan Røsæg, SalMar Farming As, Norway