When antibiotics fail on fish farms: investigate first

When a stock does not respond to an antibiotic, switching products without an investigation can delay the correct diagnosis, expose fish unnecessarily and select for resistance. In aquaculture, the prescribed dose is not necessarily the received dose. An inaccurate biomass estimate, disease-related appetite loss, uneven feed distribution, poor water conditions or a preparation error can all resemble drug failure. A second-line decision should therefore begin with observed evidence, diagnostic samples and actual exposure, not merely the medicine name.

Define failure before trying to correct it

“No improvement” only has meaning against an outcome and timeline set before treatment. Was mortality expected to decline, appetite to recover, clinical signs to stabilise or spread between units to slow? The relevant interval depends on the pathogen, fish species, life stage, temperature and lesion severity. Continued mortality during the first hours does not by itself prove that the antimicrobial is inactive.

Rebuild the episode as a timeline: onset of signs, sampling, first and last administrations, amounts offered, feed intake, daily losses, equipment incidents and simultaneous interventions. This separates a disease process that was already advanced from one progressing despite credible exposure. It also prevents an oxygen correction from being mistaken for a drug response, or an environmental deterioration from being hidden behind a change of product.

Chapter 6.2 of the World Organisation for Animal Health (WOAH) Aquatic Animal Health Code advises periodic review of farm antimicrobial records to evaluate treatment regimens. It also calls for suspected lack of efficacy to be reported to the competent authority, accompanied by susceptibility data. Reporting arrangements differ by jurisdiction, but the practical rule travels well: suspected failure is pharmacovigilance information, not only a private farm observation.

Revisit the diagnosis and protect sample quality

The first question is whether the target disease was bacterial and whether the isolated organism was truly involved. Hypoxia, ammonia, parasites, viral disease, a toxic exposure or a nutritional problem will not respond to antibiotics. A bacterium grown from a sick fish may also be secondary or incidental. Diagnosis should integrate epidemiology, clinical observations, post-mortem findings, lesions, histopathology, bacteriology and PCR when appropriate.

Before second-line treatment, sample representative fish that are untreated or least exposed whenever possible and agree the plan with the diagnostic laboratory. Tissue selection, transport temperature, time to processing and case history all affect the result. Samples collected late after several treatment days may be negative or may preferentially yield organisms that did not cause the original outbreak. That finding requires interpretation rather than an automatic “no bacteria” conclusion.

When an organism is isolated, antimicrobial susceptibility testing must use methods appropriate to aquatic bacteria. CLSI VET03 specifies media, incubation temperatures and times for different groups. Interpretive categories cannot simply be borrowed from human or terrestrial-animal testing. A minimum inhibitory concentration without a validated clinical breakpoint describes an in-vitro phenotype; it does not guarantee efficacy in the fish, at the infection site and under the stock’s environmental conditions.

Calculate what the target fish actually received

For medicated feed, four quantities must remain aligned: active substance per kilogram of fish, concentration in feed, feed allocation and treated biomass. An error in any one changes exposure. Re-estimate biomass from fish numbers, mean weight, mortality and transfers. An old inventory figure that has not been corrected can create underdosing or unnecessary waste.

Nominal feed allocation is not actual ingestion. The most affected fish often eat less, while dominant or still-healthy animals capture more pellets. Rigos and colleagues emphasise that successful antibacterial feeding depends on real feed-intake data, a credible biomass estimate and rapid, homogeneous distribution. Daily records should cover the quantity offered, refused or lost, distribution time, spatial coverage and stock behaviour.

Preparation and storage also matter. Check the veterinary product, premix concentration, batch, date, manufacturing instructions, homogeneity, water stability and storage conditions. Damaged pellets, incorrect incorporation or loss through current can turn a sound prescription into inadequate exposure. Increasing concentration on farm or extending duration does not repair those failures and may create residue and safety risks.

Examine water, the production system and the fish

Temperature influences appetite, fish metabolism, bacterial growth and pharmacokinetics. Dissolved oxygen affects both stock tolerance and feeding. Review time series rather than one convenient measurement: oxygen at critical hours, temperature, pH, ammonia, nitrite, species-relevant salinity, flow, density, solids and equipment performance.

A persistent production-system driver may sustain disease despite an active drug. Crowding, injuries, recent introductions, incomplete cleaning and cross-unit flows all deserve review. Corrective actions then belong to husbandry and biosecurity: stabilise water, separate epidemiological units, dedicate equipment and reduce infection pressure. These measures complement treatment; they do not justify rough handling or many simultaneous changes that cannot later be evaluated.

Lesion stage also limits what an antibiotic can achieve. Severely damaged gills, advanced septicaemia or profound anorexia may cause deaths even when the organism is susceptible. Separating irreversible mortality from new incident cases helps assess response. Recently dead and affected live fish usually provide more diagnostic value than a collection of decomposed carcasses.

Make a second-line decision without automatic escalation

WOAH’s practical guide for aquatic animal health professionals recommends basing second-line treatment after first-line failure on diagnostic results, including susceptibility testing. A different antimicrobial class or subclass should only be considered without test results where it is available and justified; combinations should be used only when supported by scientific evidence. This is the opposite of routine empirical escalation.

The veterinary decision should integrate the revised diagnosis, susceptibility result, probable exposure, infection site, pharmacokinetic and pharmacodynamic evidence, applicable authorisations and withdrawal period. CLSI notes that a clinical breakpoint brings together in-vitro susceptibility, exposure and clinical efficacy data. “Susceptible” therefore does not mean “effective under every condition”, while “resistant” should not be bypassed with an improvised dose.

Before restarting, define monitoring criteria: intake, mortality, signs, water parameters, adverse events and the reassessment point. Retaining laboratory reports, manufacturing records, the prescription and actual quantities delivered turns the episode into usable evidence for the farm health plan. Report suspected lack of efficacy with enough detail; certainty is not required before a suspicion becomes worth documenting.

Turn a failed course into prevention

The investigation can be organised around ten questions: was the objective defined; was the diagnosis confirmed; were samples interpretable; was susceptibility testing appropriate; was biomass accurate; was feed manufactured correctly; did target fish ingest it; were water and equipment stable; was the husbandry driver corrected; and was suspected lack of efficacy reported and reviewed? An unknown answer is a data gap to close, not a reason for a blind product change.

Vetofish can support timeline reconstruction, sampling plans, interpretation with the diagnostic laboratory, administration audits and integration of lessons into farm health and biosecurity procedures. This work supports the prescribing veterinarian and local legal framework. It cannot promise a universal antimicrobial or recovery of fish with irreversible disease.