Ehrlich's ‘magic bullet’ notion of chemotherapy foresaw substances which when given as a single dose would localize in the sites of infection and destroy the organisms there. Despite the enormous advances made in the development of antimicrobial compounds, none exhibits the remarkable properties Ehrlich visualized. Most antimicrobial agents are widely distributed in the body in response to forces that have nothing to do with infection and may result in the concentrations of the agent being least in the sites where they are most needed.
Basis for therapeutic action
Among the many properties that must be exhibited by therapeutically useful chemotherapeutic agents, the ability to achieve effective concentrations and act in the complex environment of the infected lesion is essential. When antibiotics are given systemically the delivery and maintenance of effective concentrations at the site of infection are determined by the concentrations achieved in the blood, which are in turn determined by the absorption, metabolism, and excretion of the drug and by the way in which the blood-borne drug is distributed to the tissues. From serial measurements of the concentration of the agent in the serum—usually in healthy volunteers—it is possible to calculate both its rates of absorption and elimination and the volume in which the drug is distributed. The volume of distribution indicates whether it is largely confined within the vascular compartment or spreads out into the extracellular fluid—the site of most infections—or penetrates into cells where some organisms, for example mycobacteria and brucellae, multiply. The rates of transfer, volume of distribution, and other key properties can be given numerical values which provide succinct and quantitative statements of the drug pharmacokinetics.
The time required for the concentration of drug in the plasma to fall by half is called the plasma half-life (Fig. 16.1). Half-lives of different antibiotics vary considerably. That of benzylpenicillin, for example, is only 30 min whereas that of the antimalarial mefloquine is about 3 weeks.
Fig. 16.1 Decline in plasma concentration of a drug after its intravenous injection. Three phases may be distinguished: an initial rapid fall as the drug is distributed from the plasma; a less rapid fall during the main period of excretion and metabolism; and a terminal slow decline representing, for example, the release of the drug from binding sites.
The concentration achieved in plasma while the drug is resident in the body can be measured relatively simply, but the calculation of the true half-life must take into account the distribution phase (sometimes called α-phase) during which the compound is migrating from the plasma to the tissues; this is clearly influ-enced by the route of administration, since absorption from intestinal or intramuscular sites is not instantaneous.
The half-life of a drug that is usually cited is that which follows distribution to the tissues and is designated the α-phase. Any metabolism of the drug, binding
to plasma proteins, or alteration in the functional integrity of the organs of excretion (usually the kidney or the liver, or both) will affect the elimination of the drug and hence the plasma half-life.
Drug distributed into the tissues may be bound there and this drug, as distinct from that in the plasma, may be relatively slowly remobilized and excreted. When a drug behaves in this way, the rate of elimination may be very slow and the terminal half-life (or γ-phase) very much greater than that during the β-phase.
If large doses are given or the half-life of the drug is such that complete elimination has not occurred before the next dose is administered, the concentration of drug in the plasma will progressively rise. The excretion phase being logarithmic, the rate of elimination rises as the concentration of drug rises; eventually excretion proceeds as fast as the accumulation and the drug reaches a steady state (Fig. 16.2). This may be beneficial in achieving a constant exposure of the infecting organism to the agent, but it can also be hazardous if the agent is toxic. The possibility of accumulation and its consequences must be considered when an agent with a long half-life is administered or the patient's capacity to eliminate the agent is known or thought likely to be impaired.
Fig. 16.2 Steady-state concentrations on various drug regimens. The thin lines are identical curves showing the fall in concentration of an agent with the half-life (T½) shown. The thick lines indicate the concentrations achieved by doses given at intervals of (A, C) every half-life, and (B, D) twice every half-life. In each case a steady state is eventually reached, but the time taken to achieve steady state and the concentration achieved depend on the conditions. (From O'Grady F, British Medical Bulletin 27 (1971), 142-7.)
Many antibiotics do not produce adequate plasma levels when given by mouth and are available only as injectable preparations. In some countries injections are favoured over oral therapy, but this has more to do with cultural differences and traditions than with proven therapeutic benefit. In the UK antibiotics are usually given by mouth whenever possible, particularly in domiciliary practice, because of the convenience of the oral route. In some cases, as with cefuroxime and cefuroxime axetil, considerable effort is put into the preparation of oral derivatives of the original compound. The need for properties such as stability in solution means that pharmaceutical preparations (injections, capsules, tablets, syrups, etc.) can contain different derivatives of the drug, sometimes with distinct properties.
The fraction of a dose of an oral drug that is absorbed unchanged and available to interact with the target is known as its bioavailability. The degree to which antimicrobial compounds are absorbed when given orally differs greatly. Chloramphenicol is so well absorbed that there is no place for an injectable form of the drug except in patients who cannot swallow. However, the drug is so bitter that it cannot be given as a syrup; hence a tasteless derivative is used for
such preparations which is microbiologically inactive until it is hydrolysed in the gut with the liberation into the plasma of the active form. This device of producing derivatives which have some particularly desirable property but act in the body by the liberation of the parent drug is widely used; such preparations are collectively called pro-drugs.
Some drugs that can be given orally are nevertheless irregularly absorbed and often produce low plasma concentrations. Erythromycin is one example of this, and several derivatives have been produced in an attempt to overcome the diffi-culty, including erythromycin estolate, which is microbiologically inactive, but is much more lipid soluble than the parent drug and much better absorbed in the small intestine, where non-specific esterases liberate the active erythromycin into the portal vein. Esterification as the means of improving the oral absorption of drugs has been fairly widely used, other examples being the esters of ampicillin, such as pivampicillin, and valaciclovir, the pro-drug of aciclovir.
Esters such as cefuroxime axetil allow drugs that must otherwise be administered by injection to be given orally. Others, like the erythromycin and ampicillin esters, increase the poor absorption of the native compound. These esters generally show another important benefit in addition to the enhanced plasma levels obtained: their absorption, unlike that of the parent compound, is much less affected by the presence of food in the stomach. If patients are to take drugs at regular intervals it is very difficult to separate the administration of all doses from meal times and it is a great advantage in a drug to be unaffected by such influences. A further advantage of such pro-drugs is that they are not degraded by the bowel flora and, being microbiologically inactive, the fraction of the drug that is not absorbed does not disturb the gut flora—an effect believed to lessen the chances of post-antibiotic diarrhoea and of superinfection with resistant organisms.
Interference with absorption
The original tetracycline, chlortetracycline, is a good example of an oral agent that is not well absorbed. Moreover, its naturally low absorption is further depressed by the simultaneous administration of food, especially substances such as milk that contain high concentrations of calcium or magnesium with which tetracycline forms stable insoluble chelates. The unabsorbed compound is active and disturbs the bowel flora; this most probably underlies the bowel disturbances which commonly follow its use.
Attempts in the past to minimize these side-effects by the administration of the drug with milk or alkali secured any desirable effect they may have had in limiting the gastrointestinal symptoms at the expense of drug absorption.
The most direct way of ensuring adequate concentrations of antibiotic in the blood is by intravenous injection. The highest instantaneous concentrations are, of course, achieved by a single rapid intravenous injection, but any benefit of this may be offset by rapid excretion, and many agents are given by infusion over 15–20 min. Sometimes slow infusion is necessary in order to minimize side-effects (as with vancomycin and the ‘red man syndrome’; p. 203) or local reaction at the injection site. Addition of antibiotics to drip infusions should be avoided if possible, since the slow rate of administration results in low plasma levels of the drug. Sometimes degradation of the drug in solution can occur over the prolonged period of administration, particularly if administration is combined with glucose. Certain combinations of β-lactam antibiotics and aminoglycosides mutually inactivate each other when mixed in intravenous solutions.
Rapid injection or infusion results in high concentration of the drug, which then declines rapidly as the drug diffuses into the extracellular space and into the cells if they are accessible to it. There follows a period during which the concentration of drug falls more or less rapidly, depending on the rate at which it is metabolized and excreted (Fig. 16. 1).
When all the drug is delivered directly into the plasma within a short time the plasma half-life is determined solely by the rate of elimination. When, however, uptake into the plasma is much slower then the persistence of the drug will depend not only on the rate at which it is eliminated but also on the rate at which it is added. Absorption from intramuscular sites is usually rapid, but to maintain inhibitory levels of penicillin, special depot preparations, such as procaine penicillin, have been developed (p. 18). These allow slow release from the injection site so that the plasma concentration is prolonged (Fig. 16.3).
Fig. 16.3 Blood levels produced by the intramuscular administration of (A) penicillin and (B) procaine penicillin.
Compounds in the plasma generally reach the tissues by diffusion, although in some cases there is active secretion into, for example, saliva, bile, or urine. One important factor affecting the diffusibility of compounds is the degree to which they are bound to plasma proteins, mostly albumin. With some drugs, such as cloxacillin, ceftriaxone, or fusidic acid, more than 90 per cent of the drug is bound and is antibacterially inactive. Only the diffusible fraction of the drug reaches the tissues and only this fraction exerts any antimicrobial effect. As the
unbound fraction diffuses away, more of the plasma-bound drug dissociates and the equilibrium between the bound and the free compound is maintained. Because of this effect in limiting the activity and diffusibility of the drug, high degrees of protein binding have been held to be highly disadvantageous.
However, two things have to be considered. The first is that, in order to be therapeutically effective, adequate concentrations of free drug must be achieved at the site of infection. That this occurs with the compounds mentioned, despite high degrees of protein binding, is clear from their clinical efficacy. The second is that bound drug will go wherever the protein goes and that includes the protein poured into the infected sites with inflammatory exudate. To this extent, protein-bound drug may be looked upon as a pro-drug with the valuable property of ‘homing in’ on to sites of inflammation. Once in the site, as far as is known, the concentration of free and active drug is still defined by the equilibrium between free and bound drug.
Another important aspect of protein binding is that drugs may compete for binding sites. It is possible for one drug to displace another of lower affinity, resulting in increased concentration and, possibly, toxicity of the liberated drug.
Access to infected sites
The main interest in the distribution of antimicrobial agents is in the area which is least easy to study: the concentration achieved at the infected target. Some sites can be sampled directly, important examples being the cerebrospinal fluid (CSF) or the bronchial secretions. The difficulty is that there is good reason to believe that the distribution of drug in the subarachnoid space and the bronchi
is far from homogeneous, and a comprehensive view of the behaviour of the drug in these sites is hard to obtain. What is clear is that the specialized tissues which separate the drug circulating in the plasma from these important sites of infection are normally relatively impermeable to all but a few agents. Macrolides, fluoroquinolones, and trimethoprim appear to be unusually well-distributed agents, hydrophobic or amphipathic properties apparently playing a part in this.
Most drugs enter the infected site in high concentrations through increased permeability of the vascular endothelium produced by the inflammatory process, and an important result of this is that as the drug exerts its antibacterial effect and the inflammation subsides the drug is progressively excluded again. In some cases of meningitis treated with high doses of ampicillin, for example, exclusion of the drug as inflammation subsides may lead to a recrudescence of infection.
One of the most difficult areas to study is soft-tissue infection. Some tissues, such as the tonsil, are accessible, and by homogenizing samples of such tissue in suitable fluid the ‘tissue’ concentration of antibiotics can be measured. If this concentration is significantly higher than that of the blood the result is helpful, although there is no way of knowing where precisely the drug is located in anatomical terms. If the concentration is less than that of the blood the measurements may represent simply the effect of diluting the drug present in the tissue blood vessels. This cannot be overcome by washing out the blood since this may artificially lower the tissue concentration by encouraging the diffusion of agent.
Attempts have been made to assess the antibiotic concentrations present in extracellular fluid by the creation of artificial collections of fluid which can be repeatedly sampled. In animals this has been achieved by implanting a sterile mesh into the subcutaneous tissue or muscle and allowing it to fill with fluid. In man, blisters can be raised on the skin by application of vesiculating agents such as cantharides. In general, the concentration of antibacterial agents rises and falls very much less rapidly in these spaces than it does in the plasma, and the concentrations achieved are considerably lower and more prolonged.
How closely these spaces mimic normal tissue spaces and particularly infected tissue spaces in their composition and fluid dynamics is much disputed. Interesting differences can be demonstrated in them in the behaviour of various antibacterial agents, but these do not appear to be decidedly reflected in differences of therapeutic performance.
Many antibiotics are modified in the body; the resulting metabolites are important for several reasons. Firstly, they are generally, though not always, less active microbiologically than is the parent compound. Some metabolites show not only
different degrees, but different spectra, of activity. In addition, metabolites may differ from the parent compounds in toxicity. If they are relatively inactive and more toxic, conventional microbiological assay of the drug will give very poor guidance as to the toxic hazard. Finally, the metabolites may display altered pharmacokinetic characteristics, so that the period for which they are present in the body and able to achieve an antibacterial (or toxic) effect may be longer, or shorter than that of the parent compound.
Occasionally, it is necessary to prevent metabolism from occurring. The carbapenem antibiotic imipenem is susceptible to a renal dehydropeptidase that opens the β-lactam ring. Consequently, imipenem is formulated with a dehydro-peptidase inhibitor, cilastatin, which protects it from inactivation.
Most antibiotics are eliminated by the kidneys, so that very high concentrations may be achieved in urine. Excretion is by glomerular filtration or tubular secretion, and sometimes both. Metabolites and parent drugs may be handled differently by the kidney. In pyelonephritis the principal site of infection is in the peritubular areas of the medulla. Drug can reach this space by diffusion from the plasma, but high concentrations can also be achieved by non-ionic back-diffusion. As the urine becomes more concentrated in its progress towards the collecting tubule, the concentration of drug rises markedly so that a large concentration gradient develops between the tubular contents and the interstitial space where organisms multiply. Only non-ionized species of drug are diffusible; the delivery of high tubular concentrations of the drug into the infected site therefore depends on the proportion of drug non-ionized at the pH of the tubular content. Nitrofurantoin, for example is a weak acid and in renal tubular contents the greater part of the drug will be non-ionized and free to diffuse into the peritubular space.
The principal compounds excreted in the urine by active tubular secretion are the penicillins and cephalosporins. This process is so effective as to clear the blood of most of the drug during its passage through the kidney, and these compounds generally have a very short half-life of 2 h or less. The period for which inhibitory levels of rapidly excreted agents are present in the blood can be increased by increasing the frequency of administration. Alternatively, an agent that competes for the active transport mechanism may be used. The oral uricosuric agent probenecid shares the tubular route of excretion of penicillins and can be used to prolong the plasma half-life of these antibiotics.
Biliary excretion and enterohepatic recirculation
Some antibiotics, for example erythromycin, produce very low and sometimes undetectable levels in the urine. Many such compounds are excreted in the bile,
and metabolites produced in the liver may also follow this route. When the bile enters the small intestine a significant proportion of the drug may be reabsorbed and, if so, enterohepatic recirculation can play a significant part in maintaining plasma levels of the drug. Because of its relative inaccessibility, it is difficult in practice to obtain entirely satisfactory figures for the concentration of drugs excreted in the bile. Most studies are done on bile obtained at operation or postoperatively, when the bile is draining through a T-tube and conditions are plainly not physiological. Naturally, this route of excretion ceases when the biliary tract is obstructed or hepatic function deranged, and it is plainly prudent to avoid administering to such patients agents for which this is a major route of elimination.
In contrast to pharmacokinetics, which describes the way a drug is handled in the body, pharmacodynamics specifies the way the drug interacts with the microbial target in the conditions imposed by pharmacokinetic fluctuations. This aspect of drug action is much more speculative, since it is not normally susceptible to direct measurement in situ, but is valuable insofar as it tries to define the dosing pattern likely to lead to the most efficient eradication of the offending microbe.
Among the factors to be considered are: whether the area under the concentration–time curve is more important than the peak level achieved; whether the regular replenishment of an inhibitory concentration for a short period is more, or less, efficacious than a continuously maintained inhibitory level; and whether antimicrobial activity is prolonged beyond the time for which an inhibitory concentration is achieved—the so-called post-antibiotic effect (p. 28).
Such judgements rely heavily on knowledge of how the micro-organism responds in laboratory controlled conditions in vitro and models of varying degrees of elaboration are sometimes constructed to try to approximate more closely to the dynamic circumstances that exist in life. Extrapolation of these observations to the in-vivo situation depends on the assumption that similar behaviour can be expected to occur in the complex, fluctuating conditions within an infected lesion and, indeed, on whether concentrations at the site of infection can be reliably predicted given individual variation.
Perhaps the most extraordinary feature of antimicrobial pharmacokinetics is that, after more than 50 years of intensive clinical and experimental study, the shape of drug concentration–time curve that is needed at the site of infection to secure optimum antimicrobial effects is still not known. If we knew that, we would be one step (of several required) towards more rationally based dosage schedules.