Antimicrobial Chemotherapy, 4th Edition

General principles of usage of antimicrobial agents



  1. C. B. Slack

Chemoprophylaxis is the prevention of infection by the administration of antimicrobial agents as distinct from prevention by immunization. Individuals who require prophylaxis differ from the normal population in that they are known to be exposed to a particular infectious hazard and/or their ability to respond to infection is impaired.

Prophylaxis should be confined to those periods for which the risk is greatest, so that the problems of disturbance of the normal flora, superinfection with resistant organisms, untoward reactions, and cost will be minimized. If chemo-prophylaxis is to have a place in the management of infection it is likely to be most effective when we can identify:

  • individuals in whom the risk of infection is high
  • the organisms likely to be responsible
  • the agent likely to be active against those organisms.

Failure to establish and adhere to such guidelines has made unnecessary chemo-prophylaxis the commonest form of antibiotic misuse.

Patients with normal resistance

A clear example is provided by the need to protect travellers to areas where malaria is common. The chances of acquiring the disease are high, the results can be grave, and the period at risk is well defined: from arrival in the area until 4 weeks after departure—the time taken for any parasites that may have been acquired to be finally eliminated. There are agents which are specifically active against the malaria parasite and suitable for prophylaxis, and the main complication is that the parasite may be resistant to the chosen prophylactic drug. It is plainly important to ensure that the agent used is active against the local parasite (see p. 358).


Not all situations are as clear cut as this, however, and a good example of the sort of uncertainties that arise is provided by the situation in which a child develops meningococcal meningitis. Any of the family, the school and social contacts, and the medical attendants might acquire meningococci from the child and become nasopharyngeal carriers. Some of those carriers could themselves develop meningitis. Should all receive chemoprophylaxis? Epidemiological evidence indicates that children are much more likely than adults to develop the disease and that acquisition requires close contact. The general advice, therefore, is to offer chemoprophylaxis only to close contacts and to medical attendants only if they have been exposed to unusual opportunities for transmission of the organism as, for example, during mouth-to-mouth resuscitation.

Again, the choice of agent is complicated by microbial resistance. Sulphonamides are extremely effective in prophylaxis, but sulphonamide-resistant meningococci are now common. This is particularly unfortunate because other agents such as penicillin, which are equally effective in the treatment of the disease, are much less effective (for reasons which are imperfectly understood) in prophylaxis. In the absence of information as to whether the organism is sulphonamide sensitive, ciprofloxacin or rifampicin should be used.

Surgical sepsis

The commonest circumstance in which patients with otherwise normal resistance to infection become susceptible is in the course of surgical operations. Here two situations must be distinguished: those in which the tissues involved are infected and those in which they are not.

Uninfected tissues

Here, the patient becomes liable to infection through the breach in the skin by organisms, usually staphylococci, derived from the patient's own carrier sites or from the surgical team. On the face of it, this might appear to fulfil the criteria for necessary prophylaxis: a known organism (Staphylococcus aureus) of known susceptibility (to flucloxacillin) with a risk of infection over a clearly defined period. In fact, there is no justification whatsoever for the use of chemopro-phylaxis in this way. The incidence of wound sepsis following operations on uninfected tissues, such as thyroidectomy, should be extremely small and the effect in the otherwise normal patient should in any case be easily controlled. If wound sepsis following such operations is common, something is seriously wrong with the hygienic management of the patients, and the source and mode of spread of infection must be identified and eliminated. The use of antimicrobial agents in this situation diverts attention from the real problem and, by failing to control spread, facilitates infection with resistant strains.



Infected tissues

Under this heading we can conveniently dismiss operations conducted in the presence of infection: for example, incision of an abscess, the removal of infected bone, or prostatectomy in the presence of urinary infection. Antibiotics are commonly given in conjunction with such operations, often with the specific intention of controlling any spread of organisms locally or into the bloodstream as a result of the procedure. In this sense, such therapy can be regarded as prophylaxis, but it is more appropriately regarded as part of the treatment, which will often already have begun before surgery when the causative organism and the appropriate agent will ordinarily be known from prior laboratory tests.

Contamination by normal flora

True chemoprophylaxis may be required in operations involving organs that are not infected, but which harbour a normal flora, notably the oropharnyx, the gut, and the vagina. As in the case of the skin, the oropharyngeal flora seldom gives rise to local infection following simple procedures such as tonsillectomy, and systemic chemoprophylaxis is not required. However, if there is much tissue damage or local impairment of the blood supply, such as may be seen in destructive wounds, chemically induced abortion, or elaborate vaginal repairs, implanted organisms may give rise to infection. In such cases infection often involves anaerobic organisms and several different species may act in concert to produce a so-called synergic infection. Such conditions require chemoprophylaxis and, where there is necrotic tissue, surgical debridement.

Upper gut surgery

The normal flora of the gut is uniform neither in its distribution nor in its composition. The mouth contains copious organisms, but the oesophagus, stomach, and duodenum normally contain only a few passengers derived from the mouth or the food, many of which are destroyed in the acid of the stomach and the rest hurried by peristalsis to the terminal ileum where the mixed flora typical of the large-bowel contents begins to develop. This is notable for its content of Gram-negative rods, both aerobic and anaerobic, and for the extraordinarily high concentrations of bacteria which develop as the gut contents are progressively dehydrated.

It is to be expected from the scanty flora harboured that operations on the upper gut are seldom complicated by infection and do not call for chemopro-phylaxis. Indeed, prophylactic antibiotics given to patients undergoing operations on the upper gut are likely to pave the way for the acquisition of hospital strains of bacteria that are highly resistant to antibiotics. The biliary tract is normally sterile; antibiotic administration should be required only for operations in patients with infected bile, and this is treatment, not prophylaxis. Unfortunately, the presence of bacterial infection can be determined with certainty only at the time of operation and antibiotics are, therefore, usually given routinely to patients


undergoing cholecystectomy. Because most infecting organisms are aerobic Gram-negative rods, cephalosporins are commonly used.

Large-bowel surgery

The situation in the large bowel is quite different. No other organ is so laden with bacteria able to infect normal tissues, and post-operative infection rates of 30–40 per cent are not uncommon. The problem of controlling microbial spillage from the large bowel in the course of surgery has been approached in the following ways.

Bowel sterilization

The first and original procedure was to remove the large-bowel contents as far as possible, by purging and enema. Poorly absorbed agents, such as neomycin and some sulphonamides, were (and in some places still are) widely used in conjunction with mechanical cleansing of the gut to ‘sterilize’ the bowel before operation. In fact, this failed to render the residual bowel contents sterile because of inactivity against the most numerous bowel inhabitants, the non-sporing Gram-negative anaerobic bacteria. Nevertheless, vigorous pursuit of these procedures sometimes managed to leave the large bowel sufficiently microbially denuded for the patients to become superinfected with antibiotic-resistant hospital strains of staphylococci and to suffer severe and even fatal staphylococcal enterocolitis.

Tissue sterilization

An alternative approach is to attempt to sterilize all the tissues exposed to faecal organisms by irrigation at the end of operation with a potent antiseptic or antibiotic. Considerable reduction in the incidence of post-operative sepsis has been obtained following the use of antibiotics, commonly tetracycline, or non-specific agents such as povidone-iodine. Unfortunately, absorption of drugs from large raw areas or the peritoneum is at least as rapid as from intramuscular injections and, as a result, the local concentration of agent rapidly falls and the patient is exposed to remote toxic effects of the drug just as if it had been parenterally administered. Mechanical removal of organisms by irrigation (with saline alone) before closure of the wound is effective in reducing post-operative sepsis.

Systemic chemoprophylaxis

Recognition of this problem has led to prophylaxis based on two separate procedures: elimination, as far as possible, of faecal organisms by mechanical means (possibly supplemented by oral antibacterial agents), and systemic administration of drug active against the principal potential pathogens over the period during which implantation and invasion by the organism is likely to occur. In some places, systemic antibiotics have been given for a week or two before operations—a procedure which offends against many of the rules of effective prophylaxis: it contributes nothing beyond the most recent dose to the concentration of drug in the tissues at the time of operation, it exposes patients to the risk of superinfection


with resistant organisms and to increased risk of untoward reactions, and it increases the cost.

What is required is concentrations of agent adequate to dispose of any organism likely to establish infection implanted in the tissues in the course of the operative procedure. Plainly, this does not include any period before the operation begins but, in order to secure adequate concentrations at the time of operation, the first dose of agent should be given up to an hour before. If intravenous drugs are used, these are often administered by the anaesthetist at the time of induction of anaesthesia. The question of how long tissue concentrations must be maintained cannot be answered with any certainty, but obviously must cover the time of the operation. No trial has convincingly established the period of risk, but the physiology of wound healing suggests that it is relatively short. Many regimens continue post-operative prophylaxis for 24 or 48 h. For many operations, e.g. uncomplicated appendicectomy, single-dose prophylaxis is usually adequate.

Choice of agent

There have been numerous studies comparing agents used to prevent infection after colorectal surgery. There is evidence that metronidazole, which is active only against anaerobic bacteria, is as effective as broad-spectrum agents, such as cephalosporins. However, most studies support use of a combination of agents that provide activity against aerobic and anaerobic bowel commensals and offer some antistaphylococcal activity. Expanded-spectrum cephalosporins (such as cefo-taxime) combined with metronidazole, and co-amoxiclav are widely used in the UK.

Patients with impaired resistance

What has been said so far applies to patients in whom there is no reason to suspect any impairment of the natural bacterial clearance mechanisms. There are important differences in patients with impaired resistance. Bacterial infection may be caused by a much greater variety of organisms including ‘opportunist’ commensal bacteria; the occasions of exposure may be everyday occurrences and, therefore, much less easy to identify, and the role of chemoprophylaxis in eliminating implanted bacteria is more decisive than supportive.

Impaired resistance is of two forms: local and general. Local impairment occurs where, for example, the blood supply to the area is defective or foreign bodies are present, or for some other reason the local defences are not mobilized. Important examples are amputations through the mid-thigh for obliterative arterial disease, implantation of surgical prostheses, and rheumatic endocarditis.


Surgery in obliterative arterial disease is associated with a specific serious risk. Even in the most fastidious, the perineum, upper thighs, and lower abdomen can


be soiled with gut organisms among which clostridia are common. Clostridial spores are highly resistant to antiseptics, and skin preparation before operation cannot be relied on to destroy them. Implantation of such organisms into the oxygen-deprived tissues of the ischaemic limb at amputation provides precisely the conditions most favourable to their proliferation with the production of tissue necrosis and gas gangrene. Only clostridia are responsible for this condition and all are susceptible to penicillin which, given prior to operation and for 48 h after, can be relied upon to prevent this disastrous complication and must be administered whenever such amputations are contemplated. The risk in through- or below-knee amputations is far less and penicillin prophylaxis is not mandatory. However, skin sepsis is common in this group of patients and since antiseptics remove only the most superficial organisms, peri-operative prophylaxis with agents active against staphylococci and anaerobes is commonly given.

Implantation of surgical prostheses

The deliberate introduction of foreign bodies such as cardiac valves or hip joint prostheses carries a risk of infection, the results of which can be disastrous. Chemoprophylaxis is consequently universally used, but its efficacy is almost impossible to calculate, since the incidence of infection even without chemopro-phylaxis is extremely low. A very large study mounted by the Medical Research Council and the Public Health Laboratory Service in the UK has shown a signifi-cant reduction in sepsis rates in hip replacement operations when peri-operative antibiotics were used. The study also examined the benefits of ‘clean-air’ theatres and meticulous antisepsis. The combination of such special operative techniques and antimicrobial prophylaxis reduced the infection rate to about 0.1 per cent. Bone cement impregnated with antibiotic (usually gentamicin) is widely used in joint-replacement surgery in place of costly laminar-flow operating theatres.

There is no unanimity of opinion as to which antibiotic is most appropriate when prostheses are being implanted, Moreover, the question is unlikely to be settled by clinical trial, since huge numbers of comparable patients would have to be enrolled to detect differences in the efficacy of different agents. Since staphylococci are often incriminated, the agent chosen should be active against these organisms. Many orthopaedic surgeons use cephradine or cefuroxime, which combine adequate anti-staphylococcal activity with useful cover for many Gram-negative organisms.

In cardiac surgery, broad-spectrum cover with an antistaphylococcal penicillin and an aminoglycoside has been advocated, but the common occurrence in some units of multiresistant staphylococci has led to the adoption of glycopeptides (vancomycin or teicoplanin) as prophylactic agents.

Infective endocarditis

Patients with certain congenital cardiovascular abnormalities, prosthetic heart valves, rheumatic carditis, or arteriosclerosis are subject to deposition of vegetations


on the endocardium which are liable to become infected with bacteria transiently liberated into the bloodstream. For example, salivary organisms enter the blood when a tooth is extracted, and gut organisms can enter the blood during large-bowel operations. In normal subjects these organisms are quickly cleared by the reticulo-endothelial system, but in patients with pre-existing endocardial vegetations they may become trapped and grow in the vegetation to give rise to bacterial endocarditis. A great variety of bacterial species can occasionally infect the valves in this way, but much the most important organisms are viridans streptococci from the mouth and enterococci from the gut or bladder (Chapter 23).

Identification of patients at risk

Despite the availability of many agents active against the infecting organisms, treatment of bacterial endocarditis is still far from satisfactory in terms of prolonged survival. Protection against the risks of such infection is therefore essential. The difficulty is that almost none of the requirements of soundly based chemoprophylaxis can be met in full, and because of the infrequency of infection it is impossible to measure the efficacy of chemoprophylaxis and the relative merits of competing regimens.

The best that can be done is to identify all patients with relevant heart disease undergoing all procedures likely to give rise to relevant bacteraemia, recognizing that this will fail to protect those patients who give no history of heart disease—and increasing numbers of older patients in whom the original nidus is presumably arteriosclerotic fall into this category—and those who give no history of dental or other relevant manipulations. This is an unsatisfactory situation, but the alternative is frequent prophylaxis for a much wider section of the population with all the attendant problems of emergent resistance, side-effects, and cost.

Prophylactic regimens

Thus, the choice of subjects and necessary occasions of chemoprophylaxis are hard to establish. The choice of agents and duration of treatment are scarcely less so. Since bactericidal agents are essential to eliminate the causal bacteria of endocarditis, it follows that they are essential for prophylaxis. Penicillin has long been the drug of choice in chemoprophylaxis for dental procedures where the risk is from oral streptococci which are usually penicillin sensitive.

Dentists object to regimens that require injections and are, thus, unsuited to the dental surgery and likely to add substantially to the difficulties of persuading children (and not a few adults) to undergo necessary dental treatment (itself an important part of prophylaxis against infective endocarditis). This has led to the advocacy of large single doses (3 g) of amoxycillin, which has the advantage of being well absorbed when given by mouth when it produces concentrations of drug in the blood bactericidal for the majority population of sensitive streptococci for 24 h. Where enterococci are likely to be liberated into the blood, supplementation of this treatment with an aminoglycoside is essential. In patients


allergic to penicillins, oral erythromycin (or clindamycin) or intravenous vanco-mycin can be substituted for amoxycillin. Parenteral therapy is recommended only for dental procedures carried out under general anaesthesia.

As in the case of surgical sepsis, it is not established how long bactericidal concentrations of agents must be maintained, but the episodes of bacteraemia are very short and chemoprophylaxis for more than a few hours is unlikely to be necessary. Again, it is important that the agent should be given an hour or so before operation so that the peak concentration is present in the blood at the time of maximum risk.

Long-term chemoprophylaxis

Rheumatic carditis

This is an unusual form of special susceptibility to infection in that the patients are susceptible not to the infection itself—streptococcal pharyngitis—but to remote immunological sequelae involving the heart. The chemoprophylaxis of the condition rests on unusually secure grounds: only one organism is responsible—Streptococcus pyogenes—and that has remained, to date, steadfastly sensitive to penicillin, which will eradicate the organism from the throat. Penicillin is therefore uniquely indicated. Moreover, the patients at risk are clearly iden-tified by having already suffered an attack of rheumatic endocarditis, it being well established that subsequent attacks of streptococcal pharyngitis lead to further cardiac damage. As a fresh attack of pharyngitis may occur at any time and the risk of further cardiac damage persists at least through adolescence, penicillin prophylaxis must be continuous until puberty.

Immunocompromised patients

Prolonged penicillin prophylaxis of rheumatic subjects is successful because only one organism can cause the disease, no resistant strains have emerged, and prolonged penicillin treatment is seldom complicated by side-effects. Similar circumstances exist in individuals who have had a splenectomy, either after trauma or as part of planned treatment of a haematological disease. These patients are subject to invasive disease caused by pneumococci, which, until recently, were also invariably sensitive to penicillin. As in rheumatic heart disease, penicillin is given at least until the age of adolescence, but in this case a pneumococcal vaccine is available to provide added protection.

If the susceptibility to infection of patients with rheumatic carditis or asplenia had been associated with organisms that have become successively resistant to many agents—for example Staph. aureus or Escherichia coli—frequent changes of therapy would have been required, ending almost certainly with the need to choose between abandoning prophylaxis and contemplating long-term use of


relatively toxic agents. This kind of dilemma has often to be faced in other situations where the patient's resistance to bacterial infection is reduced for prolonged periods. Examples are in patients suffering from immunodeficiency or diseases associated with impaired defences or treated with immunosuppressive agents. Such patients are subject to invasion from their own microflora and from organisms acquired in the hospital environment, including opportunist organisms such as Pseudomonas which can give rise to generalized disease. Great care is taken to deny access of hospital organisms to them; they may, for example, be physically isolated from the ward in elaborate mechanical isolators, fed sterile food and given mixtures of poorly absorbed antibiotics such as neomycin, vancomycin, bacitracin, polymyxin, and nystatin, collectively active against all the organisms, including both anaerobes and Candida, likely to enter the blood from the gut. A refinement of this blunderbuss approach is to attempt selective decontamination with two or three agents that remove most of the organisms from the oropharnyx and bowel, but preserve the anaerobic flora. This tactic is intended to discourage overgrowth of resistant organisms in the alimentary tract by a process known as colonization resistance. Selective decontamination regimens often combine oral and parenteral drugs, although systemic chemoprophylaxis has been objected to in the past on the grounds that the appropriate agents may be required for therapy if infection occurs despite the elaborate precautions.

The unconscious patient

Normally, the respiratory tract below the larynx is sterile in the sense that there is no resident flora. but organisms stray into the bronchi from the oropharynx and are removed by the mucociliary mechanism. Excess mucus is removed together with the entrapped microbes by coughing. In the unconscious patient this process is suspended and, if artificial means to maintain bronchial toilet are inadequate, multiplication of oropharyngeal organisms in the depths of the bronchi will give rise to bronchopneumonia. It is tempting to guard against this by chemo-prophylaxis directed against the common pathogenic organisms: Streptococcus pneumoniae and Haemophilus influenzae, for example, with ampicillin. In practice, the effect of such therapy is to facilitate the establishment of resistant Gram-negative rods in the oropharynx, which then promptly follow the same route into the depths of the lung where they establish an infection which is more difficult to treat and may well extend to fatal septicaemia. Several studies have confirmed the expectation that chemoprophylaxis in such circumstances increases rather than decreases the mortality from respiratory infection.

Chronic urinary tract infection

In contrast, the long-term chemoprophylaxis of intractable urinary infection in patients who cannot be controlled by short-term therapy is highly successful. Why should this form of long-term chemoprophylaxis succeed when the evidently


analogous situation in the respiratory tract is at best a failure and at worst a disaster? The reasons appear to be twofold:

  • Firstly, the agents most widely used for the purpose, nitrofurantoin and trimethoprim, do not encourage the emergence of resistant strains in the gut—from which the urinary infecting organisms are derived—in the way that resistant organisms are encouraged in the oropharynx.
  • Secondly, the drugs are excreted into the urine in very high concentrations, whereas agents used in the treatment of respiratory infection achieve concentrations in the bronchial secretions which are often considerably lower than those in the blood.

The essential ingredients of success, therefore, are unusually favourable phar-macokinetic properties and unusual failure to facilitate the emergence of resistant strains. Continuing freedom from resistance can never be relied on and there is already some suggestion that the favourable position of trimethoprim in this regard is threatened.

Urinary tract surgery

Having said that the urinary tract is unusually suited for successful long-term chemoprophylaxis, important exceptions must be noted. Unfortunately, patients are still infected in hospital in the course of instrumentation or surgery. Some of these patients have gross abnormalities of the urinary tract and may have greatly impaired defence mechanisms. Catheterization carries a significant risk of urinary infection and short-term chemoprophylaxis (for example, with nitro-furantoin) to cover the period of risk is highly successful. This success is not, however, maintained if the period of catheterization is prolonged. The defences are then set aside in a way that resembles that of the respiratory tract in the unconscious patient. Prolonged chemoprophylaxis results in superinfection with resistant organisms. This is also true of attempts to reduce infection by instillation of antiseptics into the bladder or by impregnation of catheters with antibacterial compounds. The lesson is always the same: patients in hospital, whose defences are impaired, are continuously at risk and must be continuously protected against the access of hospital organisms by appropriate hygienic measures. Attempts to prevent infection by prolonged chemoprophylaxis while hygiene is neglected can be relied upon to ensure infection with resistant organisms that may be inaccessible to treatment.

Apart from special examples in which unusual features of the organism or the agent can be readily identified, chemoprophylaxis is likely to be successful only over short periods of identifiable exposure to identifiable organisms susceptible to identifiable agents. It is likely to be worthwhile only if the risk to the patient of developing the infection it is desired to prevent clearly outweighs the possible untoward effects and cost of chemoprophylaxis.