AMDUCA allows ELU only on the order of a licensed veterinarian in the context of a valid veterinarian–client–patient relationship (VCPR). A VCPR can only exist if the veterinarian: (i) assumes the responsibility for making medical judgments about the health of the animal(s) and the need for medical treatment (and the client has agreed to follow the veterinarian’s instructions); (ii) has sufficient knowledge of the animal(s) to develop a diagnosis of the medical condition; and (iii) is readily available for follow-up in case adverse reactions or treatment failure is encountered. Such a relationship can only exist when the veterinarian has recently seen and is personally acquainted with the care of the animal(s) by virtue of examination and/or by visits to the premises where the animal(s) are kept.

ELU is limited to situations where an animal’s health is threatened or where the animal may suffer or die without treatment. Before a veterinarian can legally prescribe an approved animal or human drug one of these general conditions must also be met: (i) no animal drug is approved for the intended use; (ii) an animal drug is approved for the intended use, but the approved drug does not contain the active ingredient you require; (iii) an animal drug is approved for the intended use, but the approved drug is not in the required dosage form you require; (iv) an animal drug is approved for the intended use, but the approved drug is not in the required concentration; and (v) an animal drug is approved for the intended use, but the veterinarian has found, in the context of a valid VCPR, that the approved drug is clinically ineffective when used as labeled.

In the ornamental fish industry (nonfood-producing), veterinarians can prescribe an approved human drug for an ELU even if an approved animal drug is available if the other conditions of AMDUCA are met. Whether prescribing drugs for food-producing fish or ornamental fish, thorough recordkeeping is critical. Records of drugs used, condition treated, animal species treated, dosage administered, treatment duration, number of animals, and drug withdrawal parameters must be kept for 2 years or as required by federal or state law, whichever is greater. Thorough labeling of the drug dispensed on the veterinarian’s order is critical, including the veterinarian’s name, the name of the dispensing pharmacy, and directions for the end user similar to that described for recordkeeping.

Prior to prescription of an approved animal or human drug for an ELU in a food-producing animal, the veterinarian must: (i) make a careful diagnosis and evaluation of the condition to be treated; (ii) have an appropriate medical rationale for using a specific drug; (iii) make sure the client maintains the identity of the treated animal(s) in the record; (iv) establish a substantially extended withdrawal time supported by scientific information, and, if such information is nonexistent, take appropriate measures to assure that the animal(s) and its food products will not enter the human food supply; and (v) ensure that no illegal residues, including antimicrobial residues, occur and that the client follows the established withdrawal time before marketing food products made from treated animals.

Under the AMDUCA provisions, FDA has the right to prohibit ELU of certain drugs in animals. Currently, no approved antimicrobials are prohibited from ELU in companion animals (including aquarium or pet fish). The following drugs (both human and animal), families of drugs, and substances are prohibited for extralabel animal and human uses in food-producing animals:

• Chloramphenicol
  
• Clenbuterol
  
• Diethylstilbestrol (DES)
  
• Dimetridazole
  
• Ipronidazole and other nitroimidazoles
  
• Furazolidone and nitrofurazone
  
• Sulfonamide drugs in lactating dairy cattle (except approved use of sulfadimethoxine, sulfabromomethazine, and sulfaethoxypyridazine);
  
• Fluoroquinolones
  
• Glycopeptides
  
• Phenylbutazone in female dairy cattle 20 months of age or older
  
• Cephalosporins (not including cephapirin) in cattle, swine, chickens, or turkeys:
  
  
  
  • For disease prevention purposes;
    
  • At unapproved doses, frequencies, durations, or routes of administration; or
    
  • If the drug is not approved for that species and production class.

The following drugs, or classes of drugs, which are approved for treating or preventing influenza A, are prohibited from ELU in chickens, turkeys, and ducks:

• Adamantanes
  
• Neuraminidase inhibitors.

Injectable treatments using an antimicrobial are typically used in higher-valued fish (i.e., brood stock, some ornamental species, etc.) due to the labor and care involved in minimizing the potentially significant stress to the animal due to sedation and physical handling. This procedure can be a massive undertaking for commercial producers. Despite these challenges, advantages of administering antimicrobials by injection include assuring that all animals receive the desired dose. Oftentimes the route is intramuscular (IM), but can be intraperitoneal, intravascular, or intradorsal (caudal to the dorsal fin). IM injections are usually administered in epaxial muscles, above the lateral line and near the caudal fin. If an aminoglycoside (e.g., gentamicin) is to be used, care should be taken to inject cranial to the dorsal fin to avoid large doses entering the kidney.

Immersion or waterborne treatments are most useful when a large number of animals need to be treated, when limited stress to the animal(s) is important, and of course when it is likely that the pathogen will be exposed directly to the medicated solution (e.g., on surface of fish or gills). There are three classical ways of administering an antimicrobial via a waterborne treatment. Static baths (and dips) can be used by adding an antimicrobial solution directly to the holding system. Flush treatments are most common in flow-through systems where the entire dose is added and then the water flow is returned to slowly dilute the dose. Constant flow treatments are comprised of a constant dose injected into the water system from a stock solution.

The disadvantages of waterborne treatments include expense, waste, and potential environmental contamination. Biological filters may also be compromised due to killing the filter bacteria. A rapid rise in ammonia has been seen using therapeutic concentrations of erythromycin in a catfish recirculating system, but chloramphenicol, nifurpirinol, oxytetracycline, and sulfamerazine did not affect the filter function (Treves-Brown, 2000).

It is also important to consider the ability of an antimicrobial to be absorbed from the water. Lipophilic compounds under a molecular weight of 100 will more likely diffuse across the gills. Antimicrobials that are absorbed from the water include chloramines, dihydrostreptomycin, enrofloxacin, erythromycin, flumequine, furpyrinol, kanamycin, oxolinic acid, oxytetracycline, nifurpirinol, sulfadimethoxine, sulfadimidine, sulfamonomethoxine, sulfanilamide, sulfapyridine, sulfisomidine, and trimethoprim. Antimicrobials that are absorbed poorly or not at all include chloramphenicol and gentamicin (Treves-Brown, 2000; Reimschuessel et al., 2005).

These are a shorter and more controlled method of administering bath treatments. The advantages of this type of treatment are reduced waste (thus reduced expense) and less environmental contamination. The disadvantage of this type of approach is the increased stress to the animals through handling. Therefore most dip treatments are done when fish are small or in pet/aquarium cases, but commercial aquaculture producers have used tarpaulins to contain the antimicrobial (Vavarigos, 2003), and more recently “well boats” to more effectively contain treatments (Burka et al., 2012). Localized topical treatment, often under light anesthesia, has been recommended for small external lesions in pet fish (Stoskopf, 1993; Noga, 1996).

Some studies have used either hyperosmotic infiltration (first high osmolarity, >1200 mOsm/l, then lower osmolarity containing the antimicrobial) or ultrasound treatments to try to improve permeability across the gills. Antimicrobial absorption and elimination can be affected by salinity under normal conditions, and the effects of hyperosmotic treatments have not been adequately studied. Certain antimicrobials that bind divalent cations (such as the tetracyclines) may have their bioavailability compromised by the addition of salts. Ultrasound treatments to enhance absorption may be feasible in a small aquarium setting but have not been studied extensively. Both hyperosmotic and ultrasound treatments are fairly stressful. They are mainly used for vaccination rather than antimicrobial treatment (Treves-Brown, 2000; Navot et al., 2004).

Oral treatments are the most feasible methods for large commercial aquaculture systems because they are the least stressful for the animals; however, sick fish may not eat, resulting in lessened antimicrobial exposure and compromised therapeutic effectiveness. This was shown in a study where the concentration of oxolinic acid was examined in Atlantic salmon (Salmo salar) treated during an outbreak of winter ulcer disease (Moritella viscosa). Oxolinic acid was detected in plasma and tissues of healthy fish, whereas levels were below the limit of detection in moribund and dead fish (Coyne et al., 2004a). The moribund and dead fish also had no food in their gastrointestinal (GI) tracts. These results indicate that the antimicrobial appears to help actively feeding healthy fish fight off the infection, whereas fish with clinical signs are anorexic and therefore not receiving the antimicrobial. Variable intake can also occur if fish vary in size. Larger fish will probably consume more of the medicated feed than their smaller, less vigorous counterparts. Palatability, especially of the sulfonamide products, can also be a problem (Poe and Wilson, 1989).

Absorption from the intestinal tract may vary from species to species. As mentioned, saltwater fish will drink and, therefore, antimicrobials may bind cations in the water in their intestinal tracts, affecting bioavailability. The formulation of the antimicrobial may either enhance or decrease absorption.

Various methods for administering oral medications include commercial medicated feed, custom surface-coated feeds, custom feeds (e.g., gelatin diets), medicated live feeds (e.g., Artemia grown in or fed antimicrobials), injecting food (e.g., small fish used for food) and tube feeding (Noga, 1996; Treves-Brown, 2000). Obviously, some of these techniques are appropriate only for pet/aquarium fish.

When veterinarians decide whether or not to treat fish with an antimicrobial, they should consider the anticipated PK of the antimicrobial agent in the target fish species under the given conditions (i.e., water temperature, salinity, hardness). They should also choose an antimicrobial that is effective against the disease. A well-controlled and standardized antimicrobial susceptibility test is the best means to obtain this information. A critical component of an AST method is its ability to accurately predict a clinical outcome following treatment and/or detect emerging resistance. In other words, does a likely susceptible in vitro AST result automatically mean therapy will be efficacious? Conversely, does a likely resistant in vitro test result mean therapy will not be efficacious? The purpose of an in vitro antimicrobial susceptibility test is not to mimic in vivo conditions, but rather to provide reproducible results that can be used to predict clinical outcome.

Three reproducible AST methods are currently available for use in veterinary medicine. Agar and broth dilution susceptibility tests result in a minimal inhibitory concentration (MIC) for a single bacterial isolate and provide the most clinical relevance, where the MIC may be directly related to an achievable tissue or plasma concentration in vivo. Disk diffusion susceptibility tests yield diameters of the zone of inhibition which provide no correlation with achievable concentrations in vivo, but are simple and very affordable tests to run. The E-test has also gained popularity as a simple nonstandardized diffusion-based susceptibility test resulting in an MIC shown to yield virtually identical results as the more traditional broth microdilution tests (Luber et al., 2003).