Copyright, Purdue Research Foundation, 1996
This document contains some guidelines for antimicrobial selection and use. The points made should be considered as concepts to be used while solving problem sets.
TABLE: Facultative pathogens, common infections, and antimicrobial therapy
Name and Synonyms Usual Site of Typical Infections Effective
Infection Antibiotics
Staphylococcus Skin, upper Boils, infected Flucloxacillin,
aureus, Coagulase respiratory tract wounds, mastitis, clindamycin,
positive purulent (when sensitive,
staphylococcus parotitis, benzylpenicillin),
suppurative erythromycin,
pneumonia, vancomycin
antibiotic-induced
enterocolitis,
food poisoning,
infections
associated with
foreign bodies,
osteomyelitis
S. intermedius Skin, urinary cystitis, pyoderma amoxacillin/clavul
(dogs) tract infections anic acid,
oxacillin,
enrofloxacillin,
cephalothin
One should choose the drug that provides the best chance of success with the least liklihood of toxicity. Although that statement seems self-evident, it is surprising how many times the rule is violated. Ideally, the selection should be based on reliable clinical and bacteriological diagnosis. Note especially the word "ideally," which implies that this is not always possible. For a variety of reasons one may be forced to make a "best guess" selection of a drug. In such cases, one must make an educated guess as to the most likely organism(s) causing the infection and then, by using published tables and/or personal experience, select an antimicrobial that is effective against that pathogen. Some organisms have highly predictable sensitivity patterns whereas others do not. Tables of recommended antimicrobials for specific infections resemble those printed, in part, below and can be found in "current therapy-type" references.
For some infections, it is not reasonable to do sensitivity testing. These are typically MIXED INFECTIONS like pneumonias, polynephritides, etc.
Antimicrobial sensitivity patterns are fairly predictable for some organisms, e.g., streptococci. Tables with recommended drugs can be used for initial therapy for such organisms with some confidence. One problem with published tables is that they reflect experience in a particular locale or are the "national" experience in the past. They may not reflect current local patterns, therefore, clinicians should use recent, local experience in preference to the Tables if it is available. Again, clinical laboratory testing of the sensitivity patterns on a routine basis will provide information that can be used for initial therapy. Prophylactic therapy should also be based on local sensitivity patterns of the most likely infecting organisms, if possible.
When culture results are not available start with the best-guess as discussed above and then change after culture results are available.
For life-threatening diseases treatment should be effective against all likely pathogens until culture results available. This will usually result in the use of combinations and broad spectrum antimicrobials.
When results of bacteriological culture and sensitivities are already known, use the most effective antimicrobial with the NARROWEST spectrum. All other factors being equal, bactericidal drugs are preferrable to bacteriostatics. When body defenses are compromised, bactericidal drugs are strongly recommended.
Appropriate use of antimicrobials requires that one minimize or avoid general prophylaxis and treatment of viral infections with antibacterials. Inappropriate use and heavy appropriate use can lead to resistant strains of organisms. Thus, it is common knowledge that "hospital strains" of organisms exist that are resistant to heavily used antimicrobials. It is much better to be infected with a wild type than a hospital-acquired organism!
Timing of prophylactic therapy is also important. For example, antimicrobials should be given before surgery when there is a significant risk of bacterial contamination and it is deemed necessary to use prophylactic antibacterials. The drug must be in blood at therapeutic concentrations so it can be in the clots and other tissues entrapped during the procedure.
When the presence of an animicrobial changes the relative proportions of various species and strains of microrganisms at a location, it is said to exert selection presure. Although not often considered, some antimicrobials exert more selection pressure than others and, therefore, may be less desirable for initial, blind therapy. For example, piperacillin exerts less pressure than ampicillin. Broad spectrum antimicrobials are especially prone to exert selection pressure, hence the tendency for superinfections.
Also known as parallel resistance, cross-resistance is common among members of some antimicrobial groups. The practical importance of this is two-fold. First, if treatment with one member of the group fails it is not advisable to attempt therapy with another member of the same group if cross resistance is common in that group. Second, one can use one member of a group in sensitivity tests to predict the action of other members of the group. One then chooses one member of the group over others on criteria other than their antibacterial action, e.g., toxicity, pharmacokinetics, and cost.
Cross resistance is common among the tetracyclines; among the sulfonamides; between neomycin and kanamycin; among penicillin G, penicillin V, and ampicillin when due to penicillinase; and between ampicillin and amoxicillin. It is not common or predictable for the aminoglycosides like amikacin, gentamicin, and kanamycin.
Patient Characteristics e.g., allergies and eliminating organ function are important considerations.
Successful use of a well selected drug also depends on the dose form of the drug and how it is used. After highlighting some general pharmacokinetic considerations, the following will be briefly considered: route, doseform, dose, and regimen (dose interval and duration of therapy).
It should come as no surprise that absorption, distribution, and elimination are crucial factors that must be considered in antimicrobial therapy.
ABSORPTION RATE varies dramatically with route of administration and the formulation of the drug. As will be seen later with the ampicillin sodium and ampicillin trihydrate example, the solubility of the dose form in body fluids is extremely important. Highly soluble forms given parenterally will produce high initial concentrations in plasma, but these will fall off rapidly and dosage will have to be repeated at frequent intervals. Conversely, poorly soluble forms are absorbed slowly and produce low, but sustained concentrations. The time-course of the plasma drug concentration curve produced by the selected drug, dose form, and schedule must be tailored for the target infection and site.
DISTRIBUTION of the free (non-protein bound) drug to the tissues is a crucial factor. If the drug is limited to the circulation, then it will be of little value in infections in "protected" environments. Examples of such environments are the CNS (blood-brain barrier), the aqueous humor, testis, and mammary gland cisterns. Drugs that cross biological membranes poorly will probably not achieve therapeutic concentrations in these sites. Less obvious is the fact that an infection site may be walled-off by connective tissue or may have poor circulation. Although we may use plasma drug concentrations for therapeutic drug monitoring and as indications of probable concentrations elsewhere in the body, the astute therapist accounts for factors that lead to non-homogeneous distribution. Knowledge of these factors will be useful in selecting a drug, a dose, or a dosage regimen.
ELIMINATION of drug is also a crucial factor. Slower than expected elimination, may lead to accumulation and toxicity. Faster elimination, may result in low systemic drug concentrations leading to therapeutic failure. Conversely, rapid elimination may actually increase the concentration transiently in the effluent of the eliminating organ, e.g., urine or bile.
Route should be IV for serious infections because this is the surest way to achieve therapeutically effective systemic drug concentrations. It also offers the most control. For example, one can administer the drug continuously via an IV infusion or intermittently via an IV bolus injection. The IV regimen for serious infections may depend on whether the drug is bactericidal or bacteriostatic. The general rule of antimicrobial therapy is that the plasma drug concentration should be above the MIC for the target organism. Because this is difficult to achieve with many drugs, the effort may not be cost-effective. Thus, bolus administration may be acceptable for bactericidal drugs because of the "post-antibiotic" effect. Bacteria may be so seriously affected by the drug that they cannot recover during the time the drug concentration is below the MIC. Because the "post-antibiotic" effect is less marked for bacteriostatic drugs, they should be given by infusion for serious infections, if at all possible.
Oral (PO) administration is the easiest and most common route, except for some fractious children and animals. In some cases, it is impossible to give drugs orally, e.g., because the patient is vomiting chronically or has gastric obstruction. Gastric acidity, gut enzymes, and microflora (also rumen microflora), inactivate some drugs. Beyond these releatively constant factors, absorption is still variable with most drugs, e.g., ampicillin. But it fairly predictable with others, e.g., chloramphenicol, minocycline,and some sulfonamides.
Intramuscular (IM) and Subcutaneous (SC) administration are also widely used, but produce less predictable systemic drug concentrations than the IV route and are less practical in many cases than PO administration. Consult other sources for a complete discussion of the advantages and disadvantages of the various routes of administration for particular applications. Remember also, the impact of rates of absorption of various dose forms, bioavailability, and, in the case of meat animals, the possibility for ruining cuts with drug residues or injection damage.
Patient TOLERANCE of antimicrobials varies with route of administration. For example, topical application usually presents no problem other than allergic reactions. Oral administration may lead to gastrointestinal upset due to irritation or to alteration of the microflora. Oral administration may be less likely than IV or IM administration to produce severe allergic reactions, e.g., with the penicillins. IM injection may produce irritation and/or sterile abscesses. IV injection of highly irritating drugs may cause phlebitis unless diluted as in an infusion (e.g., amphotericin B). Conversely, IV administration is the best route for highly irritating drugs.
Topically administered drugs usually do not penetrate very well whether applied to skin or to mucous membranes. Topical applications to the nasopharynx are usually not effective. Sprays and solutions are usually better, but are less convenient than the more easily applied creams. Topical therapy should not be used with drugs that have a strong tendency to cause allergy, e.g., the penicillins.
Drug dosage depends on such things as the nature of the infection, the organism, patient characteristics, dose form of the drug, and route of administration. Dose and dose interval are highly inter-related. Dosage interval (regimen, dose schedule) is determined by the pharmacokinetics of the drug and the patient's status, the cidal/static nature of the drug, the MIC of the drug/organism pair, and the selected
dose form. The relevance of the cidal/static consideration was discussed earlier. Patient factors that lead to more rapid or slowed elimination will obviously necessitate dosage adjustment and many rationales for making these adjustments exist. An example of the relationship between dose form, dose, interval, and MIC is given below in which various regimens of ampicillin sodium and ampicillin trihydrate are compared. Concepts derived from the example are applicable to all species.
Effect of dose schedule and dose form, given IM, on plasma concentration of AMPICILLIN in mares at steady state. (Riviere/Traver 1980). NOTE: SAME TOTAL DAILY DOSE of sodium salt in panels B, C, and D! Note the dotted lines denoting the MICs for a typical E. coli and Streptococcus.
Special attention should be paid to the relationship between plasma drug concentration and MIC for the two organisms in the figures. Note the impact of dose interval on the length of time the drug concentration is below the MIC for the respective organisms. It should be noted that ampicillin, the active drug, is present at very different plasma concentrations depending on the form in which it is administered.
The dose of ampicillin trihydrate shown in the panels is recommended by the manufacturer. In many cases, often caused by FDA regulations that require titration to minimum doses, the label doses are lower than desired and intervals may be too long. The infrequent discrepancy between recommended doses/dose intervals and those deemed rational from MIC/pharmacokinetic considerations is the subject of controversy.
Drug dosage may also depend on the site of infection. Larger doses are required for infections in sites that are poorly perfused or which are poorly penetrated by drugs, e.g. CNS, testis, and eye.
Age is an important consideration in drug dosage, especially with premature babies and newborns in which biotransforming enzymes are not fully developed. Species differ in the rate at which levels of these enzymes approach those of adults. Metabolic and eliminating capacity is also decreased at the other end of life.
Therapeutic drug monitoring should be considered for antimicrobials, especially if they are toxic, e.g., like gentamicin.
How long should one treat an infection? A flippant answer is "as long as necessary." However, this is not always easily determined. For example, it is often recommended that therapy be continued for some time after disappearance of clinical signs. In other cases, the infection may not appear to be responding to therapy. Yet the drug and regimen being used may still be the best possible therapy and may eventually result in a cure.
Acute cases sometimes respond quickly, but therapy should be continued for a predetermined time based upon the infection in question anyway, especially in an immunocompromised patient. Chronic cases sometimes require protracted therapy and may not appear to be responding at first. For chronic cases that are not responding one must be certain of the diagnosis, and then make a decision as to continuing therapy with the drug. For some streptococci, there is no need to stop using penicillin G, just increase the dose and continue therapy.
Knowledge a drug is bacteriostatic or bactericidal can be used to guide drug selection for immunocompromised patients and for initial combinations of drugs for multidrug therapy of infections. The final determinant of whether a particular drug combination should be used against a specific type of infection is a well-controlled clinical trial.
The guiding principle is that one should not combine a "cidal" antibiotic with a "static" anibiotic in the therapy of the SAME PATHOGEN. The reason for this rule is demonstrated by the data in the following table in which efficacy of various drugs, alone and in combination, against Klebsiella isolates is presented.
Efficacy of Combinations Against
Klebsiella
Add-- Syner- Ant-
% %
%
Cephalothin [c]
plus --
Kanamycin [c] 9 90 0
Gentamicin [c] 5 95 0
Chloramphenicol [s]
plus --
Kanamycin [c] 26 36 38
Gentamicin [c] 13 51 36
Add = Additive Syner =
Synergistic Ant =
Antagonistic
[c] = cidal [s] = static
D'Alessandri et al., '76
It should be noted that for the particular combinations in the table on efficacy of combinations, those between cidal drugs never resulted in antagonism. In contrast, for nearly 40% of the isolates, combinations of a static and cidal drug were antagonistic. These \Iin vitro\i studies reflect what could happen in immunocompromised patients. It is difficult to demonstrate antagonism in patients with normal immune systems because the antimicrobials are only needed to buy time for the body to remove the organisms. However, when body defenses are compromised, the burden for removing the infection falls to the antimicrobial. Here antagonism between haphazardly chosen antimicrobials can be devastating and result in death. It should be noted that the use of glucocorticoids can interfere with body defenses.
Effects of antimicrobial combinations against bacterial isolates are measured by comparing concentrations of drug required to inhibit (Minimumun Inhibitory Concentration, MIC) or to kill (Minimum Bactericidal Concentration, MBC) the isolates. Typical studies are done in multi-well dishes that form a matrix. Each well is seeded with the same number of microorganisms. Concentration of the antimicrobials is varied from zero to a maximally therapeutically useful value. The bottom row of wells may contain no antimicrobial "A" and the left-most column may contain no antimicrobial "B". Thus, the lower, left-most well will have no antimicrobial and the upper, right-most well will have maximal concentrations of both. By studying the pattern of growth in the wells, one can categorize the combination as ADDITIVE, SYNERGISTIC, or ANTAGONISTIC. Additive combinations are those in which a "unit of inhibitory activity" of one drug is replaced by an equal "unit of inhibitory activity" of the other, i.e., the drugs are additive. Lines separating inhibited from non-inhited colonies across the matrix of cells tend to be linear from top left to bottom right. Organisms grow to the lower left of the line because the concentrations of the drugs are insufficient to inhibit the growth of the organisms. For synergistic drugs, the combination is much greater than would be expected by combining "killing power units" and inhibition is achieved at lower concentrations of drug. The line demarcating inhibited from non-inhibited colonies is concave downward. Most of the colonies in the matrix will be inhibited. Antagonism is the opposite and the line is concave upward, i.e., most colonies will survive. "Unit of inhibitory activity" is not a specific concentration that applies to all antimicrobials. For such comparisons the "FRACTIONAL INHIBITORY CONCENTRATION (FIC)" is used. The concentration required to inhibit an isolate is equal to an FIC of 1.0. Half that concentration is FIC 0.5, and so on. Thus, FIC is used to normalize the widely different concentrations of antimicrobials required to inhibit growth of specific pathogens so they may be compared. "FRACTIONAL BACTERICIDAL CONCENTRATION (FBC)" is a similar concept with the same uses as FIC except that the endpoint is bacterial killing.
Figures demonstrating the potential results of a combination study are presented above.
Simulated microtiter plates demonstrating three idealized results of antimicrobial interaction versus a single pathogen. The left-most column of each plate has no drug A and the bottom row has no drug B. Thus, the lower left cell of each plate has neither drug A nor B and the top right cell has maximal concentrations of each (1 FIC).
Results of studies of combinations of antimicrobials on many isolates have yielded some generalizations. These generalizations can be used for deciding which combinations are most likely to result in additive, synergistic, or antagonistic results, but only testing each drug pair against each new isolate will be definitive. Moreover, the categories have nothing to say about the wisdom of combining drugs for multiple infections or where antimicrobials have different access to the organisms based on pharmacokinetic factors. For example, one antimicrobial may be less active, but have better access, than another against an organism that characterstically has an intracellular location. Yet, addition of the drug with better activity and high extracellular concentrations, may result in improved response. This may be the case for brucellosis, which is frequently treated with a combination of an aminoglycoside and a tetracycline, a combination that violates the following rules.
Drugs are divided into two broad categories for consideration of combinations. With the caveats made previously, the guidelines/rules are as follows.
BACTERICIDAL (GROUP I) drugs that are primarily CIDAL on RAPIDLY growing organisms include penicillins, cephalosporins, aminoglycosides, bacitracin, polymyxins, nitrofurans, and colistin. Polymyxins and colistin vary from the others of this group in that they do not need growing organisms to exert their action. Thus, other agents will not interfere with their action regardless of their effect on the growth rate of the bacteria. Polymyxins and colistin are not important for treating systemic infections, today, so this observation is primarily made for the principle involved; the mechanism by which the cidal drug acts and its dependence on growth rate are important determinants of how its action is modified by combinations.
BACTERIOSTATIC (GROUP II) drugs include the tetracyclines, chloramphenicol, lincomycins, spectinomycin, macrolides (erythromycin), novobiocin, and the sulfonamides. At concentrations above those normally attained in vivo during therapy, some of these may be bactericidal. However, because such concentrations are not usually achieved during therapy, they are regarded as static.
In studies of drug combinations vs SINGLE pathogens in vitro, it has been found that SYNERGISM is relatively frequent among members of group I, but NEVER among members of Group II where activity is only additive at best.
The effect of mixing drugs from Groups I and II against a single species depends on the behavior of the organism toward the Group I drug. When the organism is RAPIDLY KILLED by the Group I drug, addition of a Group II drug usually results in antagonism.
When the pathogen is RELATIVELY RESISTANT to the Group I drug, addition of Group II drug might give synergism under some circumstances.
The effect of mixing Groups I and II antimicrobials in vivo is complex and the results predicted from in vitro studies of a specific pathogenic isolate-antimicrobial pair may not be demonstrated. Antagonism is especially difficult to demonstrate in vivo. Reasons for these discrepancies include the presence of a host's natural defenses, differential access to the organisms (pharmacokinetic differences), and multiple infectious agents.
Nevertheless, the rules are useful and antimicrobial combinations may be dangerous as the following studies demonstrate. Lepper & Dowling (1951) observed that in treating cases of bacterial meningitis, treatment with massive doses of penicillin G alone resulted in death of 30% of 43 patients. A combination of massive penicillin G and chlortetracycline resulted in death of 79% of 14 patients. Mathies et.al., (1968) made similar observations. With ampicillin alone, 4.3% of patients died whereas 10.5% died when treated with a combination of ampicillin, streptomycin, and chloramphenicol. In another study using dogs with experimental Pneumococcal meningitis, chloramphenicol was found to interfere with the action of penicillin G (Wallace et al. 1966).
A number of problems that are typically caused by misuse of antimicrobials can be avoided or minimized by observing simple precautions. Problems and misuse patterns that predispose to them are identified here.
Although no drug meets all of the criteria of the ideal agent, the construction of such a list is useful because it highlights characteristics that should be noted by therapists. The following list should be taken in that light. The drug should