Antimicrobial drug factors include mechanism of action, drug disposition, adverse effects and toxicity, and the importance of increasing resistance. All veterinarians should take particular note of the Veterinary Feed Directive and are encouraged to consult the most current FDA regulations.
Understanding the mechanism of action (MOA) of antibacterial agents is important for several reasons. The MOA determines whether the antibacterial action is likely to be bactericidal or bacteriostatic, and whether the relationship between plasma drug concentration (PDC) and organism minimum inhibitory concentration (MIC) is concentration dependent or time dependent. Further, MOA often is related to the mechanism(s) by which resistance emerges. For some antimicrobials, the MOA also relates to the mechanism(s) of toxicity. Finally, if combination therapy is to be considered, drugs with MOAs that complement one another should be chosen.
The major MOA of antimicrobial agents, with examples of each type, are as follows (see also Antibacterial Agents): 1) inhibition of cell wall synthesis: β-lactams (penicillins, cephalosporins, and cephamycins), glycopeptides (vancomycin), bacitracin, fosfomycin; 2) impairment of cell membrane function: polymyxin B, colistin; 3) inhibition of protein synthesis through binding either to a single (tetracyclines, chloramphenicol, macrolides, lincosamides) or both (aminoglycosides) ribosomal subunits; 4) inhibition of DNA synthesis and replication: novobiocin, quinolones, metronidazole; 5) inhibition of DNA-dependent RNA polymerase: rifamycins; and 6) inhibition of folic acid and consequently DNA synthesis: sulfonamides, trimethoprim.
Most mechanisms will result in killing (rather than static) efficacy, with notable exceptions being drugs that inhibit a single ribosomal unit, or the use of a sulfonamide without a diaminopyridine potentiator such as trimethoprim or ormetoprim. Knowing whether a drug is concentration or time dependent is important for design of a dosing regimen. Concentration-dependent drugs include the fluoroquinolones and aminoglycosides. They are more effective when peak drug concentrations at the site of infection exceed the MIC of the infecting microbe by 10 or more. Such drugs also exhibit a long postantibiotic effect, particularly toward gram-negative organisms. This effect results in continued inhibition of microbial growth after brief exposure to the drug. Thus, for such drugs, the dose rather than the interval, is important. For time-dependent drugs, dosing regimens should be designed such that drug concentrations remain above the MIC for most of the dosing interval. For such drugs, the dose may need to be increased to surpass the MIC of the organisms, but the interval should be designed to maintain the concentrations. Examples of time-dependent drugs include the β-lactams, "bacteriostatic" drugs, and the sulfonamides. Designing a convenient dose for time-dependent drugs with very short-half lives is difficult. For example, most penicillins have an elimination half-life of ~1–2 hr, and as such, the majority (~90%) will be eliminated in 3–6 hr (for amoxicillin, ~4 hr). As such, dosing regimens should be at least every 8, if not every 6, hr. Choosing a drug with a longer half-life, if the isolate is sufficiently susceptible, is more prudent. Some drugs are both concentration and time dependent; eg, efficacy of the fluoroquinolones can be enhanced by adding a second dose in the face of treatment of an organism with a high, yet susceptible, MIC.
Drug disposition (absorption, distribution, metabolism, and excretion) of antimicrobials can impact therapeutic efficacy and should influence the design of the dosing regimen. Most antimicrobials are administered orally. Exceptions include those destroyed by GI acidity (eg, some β-lactams, and particularly penicillins) or microbes (ie, ruminants) and those insufficiently unstable for oral preparation. Food can impact absorption of some drugs, most notably tetracyclines (except doxycycline) and the fluoroquinolones (rate but not extent impacted). A drug that has a very high oral bioavailability in one species cannot be assumed to have a similarly high oral bioavailability in another. For example, the oral bioavailability of ciprofloxacin is good in people but fair or negligible in dogs and horses. Most antimicrobials must potentially reach any body tissues. Antimicrobials that have a distribution limited to extracellular fluid, ie, "water-soluble" drugs, include the β-lactams, fosfomycin, aminoglycosides, and some members of the sulfonamides and tetracyclines. Lipid-soluble drugs distribute to total body water and include the fluoroquinolones, macrolides, clindamycin, many sulfonamides, and doxycycline or minocycline. In general, drugs that are water soluble distribute well to extracellular fluid of most organs, but there are exceptions. For example, only ~30% of amoxicillin reaches tracheobronchial secretions, and distribution into sanctuaries is limited. Doses of such drugs should be increased, particularly for susceptible isolates with MICs approaching the breakpoint. Lipid-soluble drugs, in contrast, distribute to total body water and, therefore, are better able to reach intracellular organisms or infections in sanctuaries. Some drugs are characterized by volumes of distribution that exceed 1 L/kg, indicating the drug accumulates or becomes trapped at some sites. Few antimicrobials appear to be tightly bound to plasma proteins. Exceptions include doxycycline, minocycline, and cefovecin. For cefovecin, its tight binding to circulating albumin accounts for its very long half-life. Several antimicrobials undergo hepatic metabolism, with some drugs (eg, ceftiofur, enrofloxacin) metabolized to an active metabolite that can contribute substantively to antimicrobial activity. Some drugs undergo substantive to exclusive excretion in the bile, including the macrolides, minocycline (and some of doxycycline), and clindamycin. For such drugs, care must be taken in those species that have a GI microflora subject to antimicrobial disruption; use for urinary tract infections should be avoided. Many drugs are excreted renally and subsequently concentrated in the urine, including β-lactams, aminoglycosides, most fluoroquinolones, and several tetracyclines.
Toxicity is associated with harm in the animal and thus is discriminated from adverse effects, although such effects are rarely life threatening. Because bacteria are prokaryotic, most antimicrobials do not interact with eukaryotic host targets. As such, mechanisms of antimicrobial toxicity are generally not related to mechanisms of antibacterial action. An exception is drugs that target microbial cell membranes (polymixins), the use of which is generally limited to topical administration. Additionally, mammalian mitochondrial topoisomerases are subject to damage by fluorinated quinolones, resulting in a diversity of adverse effects. Drugs that disrupt microbial flora may be lethal to some species, particularly hindgut fermenters (horses, rabbits). The release of endotoxin when treating a large, gram-negative inoculum may be problematic for some drugs (especially penicillins), particularly in horses. Several classes of antimicrobials exhibit toxicity unrelated to their (known) mechanism of action. For example, the aminoglycosides are nephrotoxic because they are actively accumulated in and eventually disrupt renal tubular cells, and selected fluoroquinolones cause retinal degeneration in cats due to phototoxicity in the retina. Chloramphenicol is associated with irreversible bone marrow suppression in people, because it is converted to a toxic metabolite. Drug interactions might contribute to some adverse effects as well. For example, chloramphenicol is a marked inhibitor of drug-metabolizing enzymes, and its use with potentially toxic drugs metabolized by the liver should be avoided. Selected fluoroquinolones inhibit selected drug-metabolizing enzymes. Macrolides as a class may compete for both P-glycoprotein and thus impact all aspects of drug movement, as well as inhibit some drug-metabolizing enzymes.
Antimicrobial resistance is a common reason for therapeutic failure. Resistance is a natural response to exposure to toxins, including antimicrobials. However, although resistance is not desirable in an infecting pathogen, resistance by itself, including multidrug resistance (MDR), is not the problem. Rather, it is the virulent organism that causes harm to the animal. Acquisition of virulence factors or genes necessary for survival in a hostile environment is responsible for illness. A virulent, MDR organism is particularly problematic.
Bacterial resistance to drugs can be either inherent or acquired. Genes coding inherent resistance are present in all strains of an organism, with expression independent of antimicrobial exposure. The “spectrum” of antimicrobial drugs reflects inherent resistance. For example, bacterial wall–defective variants (such as L-forms, spheroblasts, and protoplasts) are resistant to cell-wall inhibitors. Conversely, cell-wall access to some drugs is prevented because of impermeability reflecting very small porins in selected gram-negative bacteria (eg, Pseudomonas aeruginosa).
In contrast to inherent resistance, acquired resistance occurs only in selected strains of an organism, usually emerging in response to exposure to an antimicrobial. Genes for resistance can be acquired either spontaneously due to mutations or through sharing of genetic material via plasmids or transposons. Mutations most commonly occur by chance, with the likelihood that at least one CFU in any given population being resistant to any chosen drug increasing as the population reaches 107 CFU. Because the mutation is transmitted to daughter cells, it will remain in the population genome (“vertical” resistance). Selection pressure causes less susceptible isolates to succumb to the drug, resulting in a residual population that expresses resistance to the drug even though exposure has been discontinued. However, with time, mutated bacteria, which are often physiologically impaired, also will disappear. Generally, mutations confer resistance only to the drug or drug class, and mutated cells may be less virulent. There are exceptions, with fluorinated quinolones being an example of a mutation that emerges in the presence of the drug and targets multiple drugs. This reflects their MOA, which results in damaged DNA. Subsequent mutations include those that alter regulators of efflux pumps. A marked increase in efflux activity results in MDR.
Acquired resistance also can be shared horizontally by transfer from one organism to another. Such resistance emerges in response to the presence of the drug. Shared resistance occurs rapidly, often during the course of therapy and often resulting in the transmission of multiple genes targeting multiple drugs. Transfer via plasmids is the most recognized mechanism of shared resistance. Plasmids are composed of extrachromosomal DNA, ie, DNA not vital to cell function. Plasmids are replicons, meaning they are capable of replicating autonomously in the host. Multiple mechanisms exist whereby plasmids can enter a bacterial cell.
Transformation, which occurs in only a limited number of bacteria, is accomplished by passage of naked DNA from donor to recipient. Transduction involves transfer via a bacteriophage that inserts itself into recipient bacteria. Phage-mediated transduction occurs in some gram-positive (especially Staphylococcus aureus) as well as gram-negative species. The most common method is conjugation in which DNA passes from the donor cell to the recipient via a bridge formed during direct cell-to-cell contact. This is the most sophisticated form of transmission, because the donor must have the necessary surface appendage (sex pilus) to form the bridge that is coded for by a resistance transfer factor on the plasmid. A single CFU may contain multiple plasmids, each of which in turn may carry multiple genes conferring resistance to multiple drugs. Among the disadvantages of plasmid-mediated resistance is the ease of sharing between gram-negative organisms and less commonly between gram-positive organisms; transfer can also occur between gram-positive and gram-negative organisms. Although plasmid-mediated resistance can occur rapidly, plasmids are generally shed by the bacterium once the drug is no longer present.
Among the mechanisms whereby genetic sequences can be transferred between extrachromosomal (plasmid) and chromosomal DNA is the transposon, which is a DNA sequence that can change position in a genome. Transposons can carry chromosomal DNA from one bacterial cell to a plasmid and back. These transpositional sequences may be carried on gene cassettes. Integrons are gene-capturing systems found in plasmids, chromosomes, and transposons. Integrons can carry genes imparting antimicrobial resistance; generally, such resistance impacts multiple drugs. Once incorporated into chromosomal or plasmid DNA, the genes are subsequently expressed or disseminated even further. Problematically, in addition to resistance genes, these cassettes may also include virulence factors. This is exemplified in certain strains of methicillin-resistant Staphylococcus aureus (MRSA), which reflects acquisition of the mec gene, encoding for a mutated penicillin-binding protein-2, which prevents binding by any β-lactam drug.
Infection by MRSA has transitioned from hospital acquired, occurring only in the immunocompromised, to community acquired because of acquisition of a virulence gene that facilitates infectivity. Acquired resistance primarily reflects three major cellular mechanisms: 1) Intracellular drug concentrations can be reduced. Multiple mechanisms can accomplish this. Drug movement into the microbial cell can be prevented by decreasing porin number or size or, more commonly, by transporting the drug out of the cell through efflux transport pumps located in the cell membrane. These mechanisms result in resistance to multiple drug classes. 2) Microbes can produce enzymes that destroy certain drugs or drug classes. This mechanism generally targets only a single drug class. Examples include β-lactamases and enzymes that target fosfomycin, aminoglycosides, or phenicols. These enzymes can be expressed constitutively, or on exposure to the drug, induced such that expression is increased. In general, the addition of larger R groups on the drug molecule sterically hinders the ability of destructive enzymes from reaching the vulnerable site of the drug molecule. 3) Microbes can acquire mutations that change the target such that it no longer binds to the drug. Examples include mutated penicillin-binding proteins that confer methicillin resistance to staphylococci, mutations in DNA gyrases that confer resistance to fluoroquinolones, or mutations in ribosomal subunits that confer resistance to various ribosomal inhibitors. Mutations also generally cause single-drug resistance.
Other less common methods include the presence of alternative metabolic pathways that circumvent the effect of the drug (eg, sulfonamides) or increased synthesis of a key metabolic intermediate that would thus require higher concentrations of the drug (eg, para-aminobenzoic acid in sulfonamide resistance).
Acquired resistance will be clinically manifested as an increase in the MIC of the isolate to the drug in question. Notably, resistance can emerge in an isolate that yet remains susceptible if the increase in MIC has not exceeded the MIC breakpoints determined for that drug. For recurrent infections, identifying the underlying cause is likely to be paramount to avoiding antimicrobial resistance, including MDR. Among the clinically relevant gram-negative isolates developing MDR are E coli and Klebsiella. Gram-positive organisms developing MDR include MRSA and its canine and feline counterpart, methicillin-resistant S pseudintermedius (MRSI). Enterococcus is another gram-positive isolate for which MDR is both natural and emerging. Clostridium difficile is an example of an obligate anaerobic organism for which MDR is clinically important.
Two major aspects of antimicrobial use are of particular concern to veterinarians: the likelihood of causing a pathogenic organism to become resistant to current antimicrobial therapy and the likelihood of commensal organisms, regardless of location in the body, becoming resistant to future antimicrobial therapy. Although designing the dosing regimen such that all infecting organisms are killed will minimize the risk of acquired resistance in the patient, any and all antimicrobial use contributes to global concerns regarding antimicrobial resistance. Historically, veterinary use of antimicrobials has focused on agricultural (eg, food-animal) applications. Yet, the concern is shifting to also include use in companion animals, particularly with greater awareness of the risk of transfer of resistant commensal organisms between pets and people.
The role of antimicrobial therapy in the advent of antimicrobial resistance is well recognized. Minimizing the risk of emerging resistance might be approached based on following the “3 Ds”: decontamination, de-escalation, and designing a dosing regimen.
Decontamination decreases the spread of potentially resistant microbes and includes attentiveness to hygiene not only in the hospital but also in the home environment.
De-escalation includes avoiding inappropriate or unnecessary systemic antimicrobial use. For example, the use of low levels of antibiotics (as in animal feeds) or improper dosing regimens may lead to a high incidence of acquired resistance in a given population; as such, this practice increasingly is coming under critical scrutiny. Identifying the need to treat with systemic antimicrobials is among the most important questions to be addressed. Clinical signs consistent with infection (fever, inflammation, pain, neutrophilia) are not diagnostic of infection. The presence of bacteria does not necessarily indicate that systemic antimicrobial therapy is indicated. Bacteria may reflect normal flora, which must be distinguished from pathogenic microbes. If in a culture, this may reflect a poorly collected sample. Single rather than multiple (three or more) organisms is more indicative of a pathogen. The extent of growth may be used to support infection; for example, the presence of bacteria in urine collected by cystocentesis is not considered an infection unless >1,000–100,000 CFUs are present. Presence of infection alone does not justify systemic treatment with antimicrobials. For example, despite approval of antimicrobials to treat feline abscesses, most might be better managed by local treatment. Asymptomatic bacteriuria might be more wisely not treated with systemic antimicrobials until clinical signs become evident; nonvirulent organisms may be acting as commensals, preventing infection with more virulent pathogens. The use of topical therapy should be considered when possible; not only can much higher concentrations be achieved at the site of infection, the systemic impact of antimicrobials on microbiota can be minimized.
De-escalation might also reflect shorter duration of therapy. Identifying the most appropriate duration can be problematic. Short-term therapy at high doses and short intervals should be sufficient to kill the infecting microbe, negating the need for longer duration therapy at lower concentrations or shorter intervals that might facilitate resistance. However, for slower growing organisms or nonhealing tissues, longer durations might be prudent. The longterm use of antimicrobials for infections associated with an underlying cause is particularly problematic. Resistance is more likely in these animals, particularly if dosing regimens are (inadvertently) designed for promotion rather than for avoidance of resistance. Data are just beginning to emerge in veterinary medicine indicating that shorter duration of appropriately designed dosing regimens can be just as effective as longer-term therapy.
De-escalation also might be accomplished by initiating therapy with a higher-tier drug and switching to a lower-tier drug as soon as possible. Reasons that an antimicrobial might be considered a higher-tier drug, with use supported by culture and susceptibility data, versus a lower-tier drug, which might be used empirically, include the following: spectrum, with more narrow-spectrum drugs being reserved for problematic infection (eg, aminoglycosides and gram-negative infections, vancomycin and MRSA), safety, mechanisms of resistance that will emerge should therapy fail (eg, drugs causing emergence of extended-spectrum β-lactamases or MDR), and importance to human health (eg, vancomycin, linezolids, fosfomycin, etc).
Designing the dosing regimen should be approached such that the entire infecting inoculum is killed by the chosen antimicrobial. It may be among the most difficult “D” to implement, because it requires not only understanding the relationship between plasma or tissue drug concentrations and the MIC of the infecting microbe but also being willing to modify routine recommended dosing regimens as needed for individual patients. This requires delineation of host, microbial, and drug factors that might impact antimicrobial therapy, either negatively or positively. Even if a microbe is historically considered “susceptible,” the amount of drug required to effectively inhibit its growth is likely to be greater now than it was when the drug was originally approved. Successful treatment of infection (ie, resolution of clinical signs) does not avoid the advent of resistance. In healthy, immunocompetent animals, adequate reduction of pathogen inoculum might be sufficient for the host to overcome residual microbial growth, whereas in an animal at risk, this residual growth may emerge as a resistant population after an initial response. The more "at risk" the animal is in being unable to overcome a residual resistant inoculum, the more important that dosing regimens be designed to kill the microbes. “Dead bugs don’t mutate” should be the guiding impetus for antimicrobial use.
Designing appropriate dosing regimens requires the following: 1) The pathogen(s) should be identified and characterized, including antimicrobial susceptibility, so that the drug matches the organism as closely as possible, narrowing the spectrum of drug used. The role of culture and susceptibility testing in selecting the drug and designing the dosing regimen is becoming increasingly important. With notable exceptions, the ability to predict infecting pathogens is limited. Exceptions include respiratory tract infections in food animals, pyoderma in dogs and, with limitations, urinary tract infections in dogs and cats. However, for urinary tract infections, E coli is a cause in only 50% of cases. The more complicated the infection, the less likely it is that the infecting pathogen can be predicted. Likewise, even if the correct pathogen is identified, predicting susceptibility in all but the most uncomplicated infections is likely to be limited. This is true both historically and in the individual patient. Even historically “susceptible” organisms are characterized by higher MIC. For example, only 50% of clinical E coli collected from dogs or cats are susceptible to amoxicillin. If the animal has previously received antimicrobials, patterns of assumed susceptibility may no longer be relevant for that animal. 2) Once drugs to which the isolates are susceptible are identified, one that is more likely to penetrate the infected tissue should be chosen. This includes taking into consideration host and microbial responses to infection. Often, this requires that the drug be lipid soluble. A variety of host and microbial factors contribute to antimicrobial failure by presenting barriers to drug penetration. Debris (eg, inflammatory materials, necrotic tissue, foreign bodies) and biofilm or reduced blood flow and hypoxia contribute to failure. The organism is intracellular in location and able to avoid detrimental effects by phagocytic cells. Even urinary tract infections may present several barriers. Simply choosing a drug that is renally excreted may not be adequate. Renal function may not be normal, such that urine (and drug) is not concentrated. Bacteria are in the urine and also inside uroepithelial cells. Particularly with chronic infections, the microbes may be protected by biofilm and in a state of quiescence such that they are less susceptible to many antimicrobial drugs.
Other factors must be considered when designing the dosing regimen. Immunocompromise as a result of disease, malnutrition, or concurrent drug therapy, or local immunocompromise caused by invasive procedures may contribute to failure. The more "at risk" the animal is in being unable to overcome a residual resistant inoculum, the more important that dosing regimens be designed to kill the microbes (ie, "dead bugs don't mutate"). Because dosing regimens should be designed to kill, bactericidal drugs should be chosen whenever possible. Bactericidal drugs include β-lactams (penicillins and cephalosporins), fluoroquinolones, aminoglycosides, and potentiated (not single) sulfonamides. Tetracyclines, chloramphenicol, macrolides (eg, azithromycin), and lincosamides (eg, clindamycin) are bacteriostatic. However, the distinction between "cidal" and "static" is based on in vitro conditions and may not reflect what occurs in the animal. In general, however, it is easier to achieve killing concentrations of a bactericidal than of a bacteriostatic drug.
The use of antibiotics in food animals, including use as growth promotants, contributes to the transfer of resistance genes among bacteria and ultimately from food animals to people. Additionally, contamination of food with resistant pathogenic bacteria during the processing of food is a concern. Carcasses may be contaminated at slaughter and processing, and subsequent improper handling or cooking of the product may lead to infection in people. The development of resistant pathogenic bacteria in poultry treated with fluoroquinolones has been documented. Infection of the human population is of particular concern, because the bacterial resistance created in the animal after veterinary use of a drug or drug class may result in resistance to human drugs of the same class. Whereas the organism developing resistance might be nonpathogenic, transfer of the resistance gene to other bacteria in the human GI tract may result in a pathogenic organism becoming resistant and ultimately in therapeutic failure in people. When selecting drug therapies for food animals, veterinarians must be aware of the potential for resistance. Use of antibacterial drugs should be in the context of a valid veterinarian-client-patient relationship. Selection should be based on all information available (clinical findings, experience, laboratory data, physical examination findings, culture and sensitivity data). Pathogens should be identified, and drugs with the narrowest spectrum of activity with known effectiveness against the pathogen should be used. Client education is important to prevent unnecessary use of antibacterial agents (such as using “leftover” antibacterial drugs to treat a new occurrence of disease), to advise on proper withdrawal guidelines of any prescribed drugs, and to ensure that drugs of the proper classes are administered at proper doses and through appropriate routes.
In June 2015, the FDA released the final version of the VFD rule. The VFD amends the Animal Drug Availability Act of 1996 (21 CFR Part 558), and it delineates the FDA's strategy to promote the judicious use of antimicrobials in food animals. A VFD drug (which includes prescription drugs and over-the-counter drugs as a third potential category) is defined as "a drug intended for use in or on animal feed." The drug may be an approved new animal drug. A veterinary feed directive, or order, is the written statement issued by a licensed veterinarian, in the course of practice, that orders the use of the VFD drug, or combination VFD drug, in or on animal feed. A VFD order can only be issued under the supervision of a licensed veterinarian, and issuing veterinarians can only do so within the framework of a valid veterinary-client-patient relationship as defined by the state in which the issuing veterinarian is licensed and practicing (or federal definition in the absence of a state definition of a veterinary-client-patient relationship). The rule describes the federal process to authorize VFD drugs, which must be subject to a new animal drug approval, and to establish an antimicrobial considered medically necessary as a VFD drug. The rule further defines two categories of VFD drugs: category I, which is not associated with a withdrawal time, and category II, which requires some withdrawal time in at least one species in which the product is approved. Mixing of category II medicated feeds requires a medicated feed mill license. Guidelines by the FDA delineate the specific information that should be included on the VFD. These guidelines indicate that extra-label use of a medicated feed is not permitted.