Antimicrobial dosing strategies: Applying PK/PD principles (Proceedings)
Concerns regarding bacterial resistance to antimicrobials are increasing the awareness of rational use in human and veterinary medicine. Successful antimicrobial therapy relies on administering sufficient doses so that pathogens at the site of infection are killed or sufficiently suppressed so that they can be eliminated by the host's immune system.
Concerns regarding bacterial resistance to antimicrobials are increasing the awareness of rational use in human and veterinary medicine. Successful antimicrobial therapy relies on administering sufficient doses so that pathogens at the site of infection are killed or sufficiently suppressed so that they can be eliminated by the host's immune system. There are complex pharmacokinetic (PK) and pharmacodynamic (PD) relationships between the host, the bacteria and the antimicrobial. Pharmacokinetics studies the time course of the drug in the body; it is determined by the processes of absorption into the bloodstream, distribution to the various organs and tissues, metabolism, and finally excretion. Pharmacodynamics is the effect of the antimicrobial on the bacteria. Pharmacokinetics and pharmacodynamics are closely interrelated. PK determines the amount & timing of drug that reaches the site of action; and the intensity of a PD effect is associated with the drug concentration at that site. New information on the PK/PD relationships of veterinary pathogens and antimicrobials is rapidly emerging and changing how antimicrobials are dosed in practice.
The relationship between bacteria and an antimicrobial in the laboratory is frequently described by the minimum inhibitory concentration (MIC), which is the lowest drug concentration that inhibits bacterial growth. Infrequently determined is the minimum bactericidal concentration (MBC), which is the lowest drug concentration that kills 99.9% of bacteria. By the Clinical Laboratory Standards Institute (CLSI) definition, the MIC values are derived as serially doubling concentrations (in µg/ml). Susceptible ("S"), intermediate ("I"), and resistant ("R") designations are derived from "breakpoints" assigned by CLSI based on safely achievable plasma concentrations and results of clinical trials. When a pathogen is reported as susceptible, it means that the recommended dosage of the antimicrobial will reach plasma or tissue concentrations that will inhibit bacterial growth in vivo. When a pathogen is reported as resistant, inhibitory antimicrobial concentrations are not safely attainable in the patient. If the pathogen is reported as intermediate, then administering the antimicrobial at higher than recommended doses may result in effective therapy. The relationship between drug concentration and microbial inhibition is not a linear predication. As antimicrobial concentration increases in vitro, eventually all bacteria will be inhibited or killed.
The culture and susceptibility report is often taken at face value, but to be truly useful the report must be carefully interpreted. Antimicrobial susceptibility data does not account for:
1) Host defences: The interactions between the host and the pathogen are complex and not predicted by in vitro tests. Antimicrobial drug action takes place in concert with host defences such as humoral and cell mediated immunity, complement components, and nonspecific antibacterial factors such as lactoferrin, lactoperoxidase and lysozyme.
2) Drug distribution in the body: The "S", "I", "R" designations assigned by the microbiology laboratory are typically based on safely achievable plasma concentrations. This does not take into account for extremely high concentrations of antimicrobials achieved in organs and fluids of excretion (kidney, urine, bile) or with local administration of high drug concentrations (e.g. ophthalmic ointments, intra-articular or intraosseous implants). Pathophysiology may alter drug distribution, and some antimicrobials accumulate better in diseased tissues, such as the tetracyclines which accumulate in pneumonic lung tissues.
3) Growth rates and size of inoculum at the infection site: The incubator of the microbiology laboratory is an ideal world. Conditions are managed to promote rapid growth and rapidly dividing bacteria are very susceptible to antimicrobial drugs. Replication rates may be much slower at the infection site and MIC's are generally unreliable for slow growing bacteria. Standardized inoculums used in the laboratory may over- or under-represent pathogen numbers in infected tissues.
4) Mixed infections: Separate susceptibility testing of pathogens in a mixed infection does not account for the pathological synergism between bacteria. The by-products of one bacteria species may facilitate the establishment and growth of another.
5) Infection environment: Many antimicrobials are inactive in purulent exudate, which is typically anaerobic, acidic and hyperosmolar. Some antimicrobials will have different activity in body fluids (plasma, milk, bile) than in nutrient-rich laboratory media. Deposition of fibrin may alter tissue penetration of antimicrobials. Many bacteria are capable of producing a polysaccharide slime capsule to protect them from host factors.
6) In vivo synergism may occur with antimicrobial combinations: Despite predictions of resistance from susceptibility testing, therapy may be successful because of synergistic combinations of antimicrobials. Synergism between penicillins and aminoglycosides has been recognized for streptococcal, enterococcal and staphylococcal infections. The synergism is attributed to increased cellular uptake of the aminoglycoside after cell wall damage from the penicillin. Likewise, antagonism may occur between two classes of antimicrobials, such as macrolides and chloramphenicol which both compete for the same bacterial 50S ribosomal binding sites. Remember however, in vitro antimicrobial synergism or antagonism does not mean it will happen in vivo!
7) Topically administered antimicrobials are often not tested: Veterinary microbiology laboratories may not routinely do susceptibility testing for antimicrobials that are only used topically. Polymixin B is one of the most effective antimicrobials for superficial Pseudomonas infections, but it causes neurotoxicity and nephrotoxicity if administered systemically, so it is rarely included in susceptibility testing. Bacitracin and mupirocin are other examples of topical antimicrobials are rarely tested for in diagnostic laboratories. Bacitracin is used topically for its Gram-positive activity, as penicillins and cephalosporins are not available in topical formulations because of the risk of sensitization and subsequent fatal anaphylaxis. Mupirocin is a very effective therapy for cutaneous staphylococcal infections, but it is not used systemically.
8) The CLSI breakpoints may be inappropriate: The CLSI breakpoints were originally established with bacterial isolates from humans, using human pharmacokinetic data and clinical trials in humans. A veterinary subcommittee was only established in 1993 and it has only recently proposed veterinary-specific guidelines for susceptibility tests for some antimicrobials.
Therefore, the true relevance of any in vitro MIC predicting the in vivo results of drug therapy is questionable. But by convention, drug dosage regimens use a target plasma drug concentration that is based on some multiple (2 to 10, most often 4) of the in vitro MIC.
A successful antimicrobial dosage regimen depends on both a measure of drug exposure (pharmacokinetics, PK) and a measure of the potency of the drug against the infecting organism (pharmacodynamics, PD). New information is rapidly emerging regarding the PK/PD relationships that determine antimicrobial efficacy in both human and veterinary patients. The PK parameters used in drug dosage design are the area under the plasma concentration versus time curve (AUC) from 0 to 24 hours, the maximum plasma concentration (Cmax), and the time the antimicrobial concentration exceeds a defined PD threshold (T). The most commonly used PD parameter is the bacterial minimum inhibitory concentration (MIC). In relating the PK and PD parameters to clinical efficacy, antimicrobial drug action is classified as either concentration-dependent or time-dependent.
For antimicrobials whose efficacy is concentration-dependent, high plasma concentrations relative to the MIC of the pathogen (Cmax:MIC), and the area under the plasma concentration-time curve that is above the bacterial MIC during the dosage interval (area under the inhibitory curve, AUC0-24hr:MIC), are the major determinants of clinical efficacy. These drugs also have prolonged post-antibiotic effects (PAEs), thereby allowing once a day dosing while maintaining maximum clinical efficacy. For fluoroquinolones (enrofloxacin, orbifloxacin, difloxacin, marbofloxacin), clinical efficacy is associated with achieving either a AUC0-24hr:MIC >125 or a Cmax:MIC>10. For aminoglycosides (gentamicin, amikacin), achieving a Cmax:MIC>10 is considered optimal for efficacy. Other antimicrobials that appear to have concentration-dependent activity include metronidazole (Cmax:MIC>10-25) and azithromycin (AUC0-24hr:MIC>25). For some pathogens with very high MIC values, such as Pseudomonas aeruginosa, achieving the optimum PK/PD ratios may be impossible with label or even higher than label dosages. In such cases, under dosing is ineffective and merely contributes to antimicrobial resistance.
For antimicrobials whose efficacy is time-dependent, the time during which the antimicrobial concentration exceeds the MIC of the pathogen determines clinical efficacy (T>MIC). How much above the MIC and for what percentage of the dosing interval concentrations should be above the MIC is still being debated and is likely specific for individual bacteria-drug combinations. Typically, exceeding the MIC by 1-5 multiples for between 40-100% of the dosage interval is appropriate for time-dependent killers. The T>MIC should be closer to 100% for bacteriostatic antimicrobials and for patients that are immunosuppressed. So these drugs typically require frequent dosing or constant rate infusions for appropriate therapy. In sequestered infections, penetration of the antimicrobial to the site of infection may require high plasma concentrations in order to achieve a sufficient concentration gradient. In such cases, the AUC0-24 hr:MIC and/or Cmax:MIC may also be important in determining efficacy of otherwise time-dependent antimicrobials. The penicillins, cephalosporins, most macrolides and lincosamides, tetracyclines, chloramphenicol and the potentiated sulfonamides are considered time-dependent antimicrobials.
When designing specific antimicrobial dosage regimens, a specific plasma drug concentration is targeted. High plasma antimicrobial concentrations are assumed to be advantageous in that a large concentration of drug will diffuse into various tissues and body fluids. Utilizing the previous information, antimicrobial dosage regimens are designed in one of two ways; either to maximize plasma concentration or to provide a plasma concentration above the bacterial MIC for some percentage of the dosage interval. For concentration-dependent killers with a prolonged PAE whose PK/PD relationship is to have an ideal Cmax:MIC, and if the volume of distribution (Vd) of the antimicrobial is known, a precise drug dosage regimen for the pathogen can be calculated from the following equation:
Dose = (Vd) X (desired plasma concentration)
Where the desired plasma concentration is some multiple of the MIC (usually 8-10) and once a day dosing is assumed.
For concentration-dependent killers whose PK/PD relationship is to have an ideal AUC0-24hr:MIC, the following equation can be used to calculate a dose per day:
Dose = (AUC0-24hr:MIC) X (MIC) X (Cl)
F X 24 hr
Where AUC0-24hr:MIC is ≥ 100, Cl is clearance (volume of blood cleared of drug per day in ml/kg/day) and F is bioavailability.
For time-dependent killers, the objective is to keep the average plasma drug concentration above the pathogen's MIC for the duration of the dosage interval. Again, utilizing Vd and elimination half-life information, you can precisely calculate a dosing regimen:
Dose = (desired ave plasma conc) X (Vd) X (dosage interval)
1.44 X T½elim
For veterinary practitioners, using these equations requires knowing where to find out all the PK and PD parameters. For example, if your diagnostic lab does not report MICs but only lists S, I, or R; you will have to guesstimate the MIC. A quick PubMed search can give you a good idea of the typical MIC range for common pathogens, but remember that this data may not be applicable for your particular geography. Likewise, there are many sources for finding the typical AUC, Cmax, half life, bioavailability, clearance, and Vd values for common veterinary drugs. Plumb's Veterinary Drug Handbook, VIN, and the Compendium of Veterinary Products are good place to start. Your drug reps also have access to a wealth of data used for the drug approval process, so contact them if you can't find the particular information you need. One other note: the PK and PD parameter values may vary widely between different studies, giving you a wide dosage range. Remember that this PK/PD information is meant to help you design and optimize your dosing regimen, but at the end of the day you will still need to use your best judgement!
Craig WA. Interrelationship between pharmacokinetics and pharmacodynamics in determining dosage regimens for broad-spectrum cephalosporins. Diagnostic Microbiology and Infectious Disease 1995;22:89-96.
Gunderson BW, Ross GH, Ibrahim KH, et al. What Do We Really Know About Antibiotic Pharmacodynamics? Pharmacotherapy 2001;21:302S-318S.
Lees P, Cunningham FM, Elliott J. Principles of pharmacodynamics and their applications in veterinary pharmacology. J Vet Pharmacol Ther 2004;27:397-414.
McKellar QA, Sanchez Bruni SF, Jones DG. Pharmacokinetic/pharmacodynamic relationships of antimicrobial drugs used in veterinary medicine. Journal of Veterinary Pharmacology and Therapeutics 2004;27:503-514.
Toutain PL, Lees P. Integration and modelling of pharmacokinetic and pharmacodynamic data to optimize dosage regimens in veterinary medicine. Journal of Veterinary Pharmacology and Therapeutics 2004;27:467-477.
Giguere S, Prescott JF, Baggot JD, et al. Antimicrobial Therapy in Veterinary Medicine, 4th Ed. 2006: Blackwell Publishing.