Antimicrobial agents are among the most commonly prescribed drugs. Their use has had a major impact on the control of most bacterial infections in man and to a lesser, although constantly increasing, degree, is affecting the outcome of many fungal, viral, protozoal, and helminthic infections. However, there are concerns that unnecessary use is compromising their beneficial effect. The principles governing the use of antimicrobial agents to be discussed in this chapter apply specifically to the management of bacterial infections, although the overall approach is similar when selecting treatment for other microbial diseases.
Antimicrobial therapy demands an initial clinical evaluation of the nature and extent of the infective process and knowledge of the likely causative pathogen(s). This assessment should be supported, whenever practicable, by laboratory investigation aimed at establishing the microbial aetiology and its susceptibility to antimicrobial agents appropriate for the treatment of the infection. The choice of drug, its dose, route, and frequency of administration are also dependent upon an appreciation of the pharmacological and pharmacokinetic features of a particular agent. Furthermore, the range and predictability of adverse reactions of a particular compound should be kept in mind.
The clinical evaluation should define the anatomical location and severity of the infective process. The history and examination frequently determine such infective states as meningitis, arthritis, pneumonia, and cellulitis. Although such diseases may be caused by a wide variety of organisms, the range of pathogens is usually limited, and the pattern of susceptibility reasonably predictable. This, therefore, permits a rational selection of chemotherapy in the initial management of such infections.
The anatomical location is not only critical from the point of view of the most likely pathogen and the most suitable choice of drug, but also determines the route of administration. Superficial infections of the skin, such as impetigo, which is caused by Streptococcus pyogenes, or infection of the mucous membranes such as oral or vaginal candidiasis, caused by Candida albicans, respond well to topical application. However, if infection is caused by the microbial invasion of tissues or the bloodstream, adequate tissue concentrations of a drug may be achieved only by intravenous or intramuscular administration.
Other clues as to the nature of the infection are gleaned from epidemiological considerations such as the age, sex, and occupation of the patient. In tropical countries, diseases such as malaria, amoebiasis, and salmonellosis (including typhoid fever) are prime suspects in the investigation of fever and diarrhoea, and local knowledge about the prevalence of diseases such as filariasis, schistosomiasis, and trypanosomiasis, which are circumscribed in distribution, may be used to advantage. In countries free from these diseases as indigenous problems, a history of overseas travel should alert the physician to consider exotic infections.
Pre-existing medical problems may predispose to infection; such conditions include valvular heart disease, underlying malignant disease, or the presence of prosthetic devices such as artificial hip joints, heart valves, or intravascular cannulae.
Under some circumstances the invading pathogen may be part of the host's normal flora. The normal host defences may be breached in a variety of ways. For example, the skin or mucous membranes, which are normally a most effective barrier against infection, may permit access of pathogenic organisms to the deeper tissues when traumatized by a surgical incision or by accident. Similarly, burns can denude large areas of the body with subsequent infection by bacteria, notably Pseudomonas aeruginosa, Staphylococcus aureus, and Str. pyogenes,which may be acquired from contact with patients or staff within the hospital.
The circulating and tissue phagocytes, together with the complement system and antibodies, provide an important defence against infection. Therefore, an absolute or relative deficiency of circulating polymorphonuclear leucocytes is commonly associated with recurrent, frequently serious, infection. In patients with acute leukaemia, cytotoxic chemotherapy often depresses the circulating leucocytes to low levels for several days or weeks. Such patients are extremely vulnerable to serious episodes of infection, particularly bloodstream invasion, which carries a high mortality if untreated.
Few infective conditions present such a typical picture that a definitive clinical and microbiological diagnosis can be made without recourse to the laboratory. Therefore, whenever possible, a clinical diagnosis should be supported by laboratory confirmation. Such confirmation makes both the diagnosis and the management, in particular the selection of antimicrobial chemotherapy, more certain and allows for a more sound assessment of the likely prognosis. However, when infection is obvious or strongly suspected on clinical grounds, therapy should be instituted as soon as appropriate specimens for laboratory investigation have been taken. In some cases (e.g. pneumococcal or meningococcal meningitis) the patient's chances of survival are directly related to the promptness with which therapy is started. Furthermore, laboratory reports are not always contributory and several days may be lost trying to establish a microbiological diagnosis, during which time the patient's condition may deteriorate.
Serological tests that demonstrate antibody against specific microbial antigens are important in the diagnosis of more persistent infections such as syphilis, brucellosis, and Q fever. Tests to demonstrate the presence of microbial antigens are also valuable in the diagnosis of selected infections. For example, fluorescent antibody reagents can detect Pneumocystis carinii (jiroveci) in sputum or bronchial lavage material, and pneumococcal antigen is often present in the urine of patients with pneumococcal pneumonia.
Assessment of sepsis and the systemic inflammatory response
Sepsis is defined as the combination of symptoms or signs of a localized primary site of infection plus a systemic inflammatory response. The presence of systemic inflammatory response is often the first sign that infection is spreading from the primary site and that the patient may be bacteraemic (see Chapter 22). The systemic inflammatory response syndrome is defined by the presence of two or more of the following indicators:
Infection is not the only cause of systemic inflammatory response syndrome. Other common causes include: accidental or elective trauma (it is a normal reaction to elective surgery); chronic inflammatory conditions (e.g. arteritis, systemic lupus erythematosus); and malignancy (especially lymphoma but also solid tumours). Also, the syndrome can be a response to infection by any type of pathogen, bacterial, fungal, protozoal, and viral. None the less measurement of inflammatory response is a key part of the clinical assessment of bacterial infection. For example, if a woman presenting with dysuria plus frequency is found to have the systemic inflammatory response syndrome it means that she is unlikely to have simple cystitis (Chapter 20), but that infection has spread into the kidney and possibly into the bloodstream.
Severe sepsis is defined as sepsis plus evidence of organ dysfunction, hypoperfusion or hypotension. Evidence of perfusion abnormalities affecting the vital organs (brain, heart, kidneys, lungs) includes acute confusion, hypotension, oliguria, and hypoxia or lactic acidosis.
Septic shock is defined as sepsis with hypotension that persists despite adequate fluid resuscitation, along with the presence of perfusion abnormalities.
The importance of this classification of the severity of systemic infection is clear from considering 30-day mortality associated with bacteraemia. On average this is 10-20%, but increases to 20-30% with systemic inflammatory response syndrome, 30-50% with severe sepsis and 50-80% with septic shock.
Selection of antimicrobial chemotherapy
In-vitro testing of drugs provides indirect evidence of the likely clinical response of a particular pathogen to a specific drug or drugs. Confirmation of clinical efficacy can be determined only in vivo; hence the importance of clinical evaluation of all new antimicrobial agents. Controlled experimental evidence gained from the treatment of artificial infections in animals provides only indirect evidence of the likely clinical efficacy. Occasionally in-vitro evidence of activity is not borne out by in-vivo evidence of success. For example, Salmonella entericaserotype Typhi is susceptible in vitro to many drugs active against Gram-negative bacilli, including gentamicin; however, typhoid fever responds clinically only to a limited range of drugs, including ciprofloxacin, ceftriaxone, chloramphenicol, amoxicillin, and co-trimoxazole. This may in part be due to the intracellular location of the pathogen in this disease.
Bacteristatic or bactericidal agents
Antibacterial agents are often separated into either bactericidal or bacteristatic agents according to their ability to kill or inhibit bacterial growth.
Examples of bactericidal drugs include the β-lactam agents and fluoroquinolones; bacteristatic agents include the tetracyclines and chloramphenicol. This separation is somewhat artificial since some bacteristatic drugs may be bactericidal either in higher concentrations or against different bacterial species. Bacteristatic agents must rely on host defences, in particular the phagocytic cells, to finally eliminate the infection, since if the drug is withdrawn bacteria have the opportunity to recover. Under most circumstances the choice between a bactericidal or a bacteristatic agent is not critical. This is not the case in the treatment of infective endocarditis. Here bacteria are protected against phagocytic activity within the vegetations present on the deformed or prosthetic heart valve or adjacent endocardium. Under these circumstances it is important to use a bactericidal drug or combination of drugs that penetrate the vegetations and thus eradicate the infection. Similarly, patients with neutropenia from cytotoxic chemotherapy or other causes of bone marrow aplasia are extremely vulnerable to infection. Bacteristatic drugs are inappropriate in these cases and bactericidal agents should be selected.
The aim of chemotherapy is to eliminate an infection as rapidly as possible. To achieve this a sufficient concentration of the drug or drugs selected must reach the site of infection. The choice of agent is, therefore, as much dependent upon the pharmacological and pharmacokinetic features of the drugs, which determine absorption, distribution, metabolism, and excretion, as upon its antimicrobial properties. These aspects are discussed in more detail in Chapter 14.
In general, drugs are administered either topically, by mouth, or by intravenous or intramuscular injection. Oral absorption is most erratic. Drugs must first negotiate the acid condition of the stomach before being absorbed, usually from the proximal small bowel. This occurs most readily when the stomach is empty and it is generally advised that they be swallowed approximately 30 min before or 4 h after a meal.
Absorption can be increased by protecting a drug from acid inactivation by a coating (so-called enteric coating), which subsequently breaks down once the tablet is beyond the stomach. Alternatively, the drug may be modified chemically to produce a more acid-stable formulation (see p. 198). For most minor infections, including skin, soft-tissue, respiratory tract, and lower urinary tract infection, oral therapy is appropriate.
In contrast to oral administration, intravenous administration avoids the vagaries of gastrointestinal absorption, and achieves rapid therapeutic blood and tissue concentrations. Intramuscular administration requires absorption through the tissue capillaries and is generally rapid except in conditions of cardiovascular collapse and shock, when tissue perfusion is impaired. Relatively avascular sites such as the aqueous, and in particular the vitreous humour of the eye, are difficult sites in which to achieve adequate concentrations of drugs. In contrast, the presence of inflammation increases the permeability of many natural barriers such as the meninges and in this situation allows higher concentrations of certain drugs, such as the penicillins, to be achieved within the cerebrospinal fluid. Other drugs, most notably chloramphenicol, are little influenced by such inflammatory changes.
Choice of antimicrobial regimens
There are no universally applicable guidelines for drug dosing. However, awareness of the relationship between the pharmacokinetic profile of a drug and the minimum inhibitory concentration (MIC) against a target pathogen has greatly assisted in better defining dosage regimens of some antibiotics. These provide a pharmacodynamic prediction of the optimum dosage regimen. Other factors that influence drug dosing are tolerability and toxicity. Some agents, most notably the penicillins, have such a wide margin of safety that high doses are frequently prescribed. Only in a few cases (e.g. treatment of Ps. aeruginosa infection with ticarcillin) does such antimicrobial overkill have a microbiologically rational basis. The ratio of peak plasma concentration (Cmax) to the MIC, or area under the curve (AUC) to MIC (Fig. 13.1) is often used to calculate the dose and frequency of administration to ensure the most effective drug concentrations. The Cmax:MIC ratio is the best predictor of bacterial killing for the aminoglycosides and quinolones, while the AUC:MIC ratio is used to determine dosage schedules for β-lactam agents. In the case of bacterial meningitis much higher ratios are required to achieve therapeutic concentrations in the cerebrospinal fluid. However, for many licensed agents, the dosages are based on experience gained from clinical trials of the treatment of a wide variety of infections and are included in the Summary of Product Characteristics Data Sheets for approved agents.
Fig. 13.1 Relationship between the pharmacokinetic profile of an antibiotic and the minimum inhibitory concentration against a hypothetical target micro-organism. (A) Minimum inhibitory concentration; (B) time above minimum inhibitory concentration; (C) peak; (D) area under the curve > minimum inhibitory concentration. Redrawn from: Finch RG. Antimicrobial therapy: principles of use.Medicine 2005; 33: 42-46 with permission from Lippincott, Williams & Wilkins.
Length of therapy
Treatment should continue until all micro-organisms are eliminated from the tissues or the infection has been sufficiently controlled for the normal host defences to eradicate it. This end-point is in general determined by clinical observation and evidence of the resolution of the inflammatory process such as the return of body temperature and white cell count to normal. Microbiological end-points may be appropriate for some infections, such as urinary tract infections where repeat urine cultures can be obtained 1 week and 4 weeks after stopping therapy. These will indicate failed treatment and recurrent infection respectively. However, other outcomes include length of stay and time to discharge (for those in hospital), and time to return to former activities. These factors are important when determining the health economics of disease management.
Many infections come under control within a few days and 5-7 days' treatment is often sufficient. Uncomplicated urinary tract infections usually respond very rapidly to chemotherapy. Selection of the least dose compatible with complete resolution is desirable (see Chapter 20). In contrast, patients with pulmonary tuberculosis require 6 months' treatment with isoniazid and rifampicin if relapse is to be prevented (seeChapter 25). Furthermore, 10 days' penicillin treatment is necessary to eradicate Str. pyogenes from the throat in patients with streptococcal tonsillitis, although symptomatic improvement occurs within a few days. There is no universally ‘correct’ duration of chemotherapy and each problem should be judged on its merits based on the clinical response to treatment.
Antimicrobial agents, like all other therapeutic substances, have the potential to produce adverse reactions. These vary widely in their nature, frequency, and severity. Many reactions, such as gastrointestinal intolerance, are minor and short lived but others may be serious, life-threatening, and occasionally fatal. Drug reactions are unfortunately a common cause of prolonged stay in hospital or may precipitate hospital admission. Drug reactions may be predictable and dose dependent, for example nephrotoxicity associated with the use of the antifungal agent amphotericin B. However, many adverse reactions are unpredictable. The subject is discussed more fully in Chapter 16.
In general, single-drug therapy of established infections is preferred, whenever possible. Such an approach is known to be effective and reduces the risks of adverse reactions and drug interactions that may accompany multiple-drug prescribing, as well as the cost of treatment. None the less, there are a few situations in which combined chemotherapy has definite advantages over single drug therapy.
In the management of acute and potentially life-threatening infections, combined chemotherapy covering all likely pathogens is often used until the cause of the infection is established. It is common practice to combine flucloxacillin with an aminoglycoside, such as gentamicin, in the initial treatment of serious infections. However, should there be evidence that the infection has arisen in association with mucosal surfaces, such as the gut or female genital tract, then metronidazole is frequently added to meet the possibility of a mixed anaerobic and aerobic bacterial infection. Once a definitive diagnosis is established it is important to adjust the therapeutic regimen to one that is most appropriate.
Under some circumstances combined chemotherapy is selected for its known synergic effect on a pathogenic organism. This increased ability to inhibit or kill the pathogen may speed resolution or reduce the risk of relapse when treating difficult infections. One of the commonest requirements for synergic therapy is the treatment of infective endocarditis caused by enterococci and occasionally by oral streptococci. The combination of two bactericidal drugs, penicillin and gentamicin (or streptomycin), is synergic both in vitro and in vivo and is associated with a more favourable response to treatment than is single-drug therapy.
Some drugs may have an opposite effect and be antagonistic. For example, shortly after penicillin and tetracycline became available, it was shown that the two drugs together produced a worse clinical result in the treatment of pneumococcal meningitis than did either drug alone. In this situation a bacteristatic agent (tetracycline) prevents penicillin from achieving its bactericidal effect on the cell wall, which is dependent on bacterial growth.
Prevention of drug resistance
It is uncommon for the bacterium causative of an infection to become resistant during treatment, although some drugs—including rifampicin, fusidic acid, and nalidixic acid—encourage the rapid emergence of resistant bacteria. These bacteria do not develop resistance in response to treatment, but small numbers of pre-existing resistant mutants proliferate when the sensitive population is suppressed. Mutational resistance is more likely to arise in severe infections where the burden of infecting micro-organisms is high, such as accompanies severe multilobar pneumonia. In the treatment of tuberculosis combined chemotherapy is used specifically to prevent the emergence of resistant variants present in the tuberculous tissues (Chapter 25).
Antimicrobial drugs vary widely in their cost. In general, generic drugs cost less than proprietary preparations, and well-established, widely used agents tend to be less expensive. Injectable preparations are usually more expensive than oral preparations, and syrups and drops are usually more expensive than tablets and capsules.
The use of very high dosage also escalates the cost. Sometimes this is unavoidable, as in the treatment of Ps. aeruginosa infections with large doses of expensive antipseudomonal agents, but, in general, high dosage should not be used without justification.
Among the most expensive antimicrobial agents are the parenteral cephalosporins, macrolides, and quinolones, all antipseudomonal compounds, lipid formulations of amphotericin B, and most antiviral agents. It is, therefore, apparent that whenever drugs that are both equally effective and tolerated are available then it is reasonable to select the cheaper agent. This is a most important consideration in some developing countries where drug costs can account for 20% of an already meagre health budget. The British National Formularyprovides a useful guide to the cost of drugs for doctors in the UK.
In addition to the unit cost of a drug, other direct costs include the use of disposable materials for drug administration, staff time in preparation and administration, and assay costs to ensure adequate and safe dosaging.
Comparative cure and relapse rates are also under constant scrutiny since the need for further treatment entails additional expense for the health service and more importantly risks for the patient. Economic evaluation is a key part of the assessment of healthcare but should not be limited to costs; a full economic evaluation must include measures of outcome for the patient and others such as family members or carers. The application of economic theory to evaluation of medicines is called pharmacoeconomics.
Failure of antimicrobial chemotherapy
Patients with established infection may fail to respond to antimicrobial therapy for a variety of reasons.
Choice of therapy
First, the choice of drug may be inappropriate for the infecting strain. This stresses the need to establish a microbiological diagnosis whenever possible so that the in-vitro susceptibility of the pathogen can be confirmed. On the other hand, some infections are caused by intracellular pathogens, as in brucellosis, chlamydial disease, legionella infection, and typhoid fever. For treatment to be successful sufficient antibiotic must penetrate the cell; this limits the number of agents that are clinically effective. Failure may also result if the drug is inadequately concentrated at the site of the infection; this may occur if the dose is insufficient or the route inappropriate. By changing to the parenteral route or increasing the dose, therapeutic success may follow.
The route of excretion also requires consideration. For instance, in renal failure, drugs normally excreted by the kidneys may fail to reach therapeutic concentrations in the urine so that treatment of a urinary tract infection with drugs such as nalidixic acid and nitrofurantoin may be unsuccessful.
Presence of necrotic material
Treatment may also fail because of the presence of necrotic material, an eschar, or abscess. Antibiotic penetration of such avascular material is poor, so surgical debridement of necrotic material or drainage of pus should be carried out early. Antibiotic treatment under these circumstances is an adjunct to such surgical management.
Presence of foreign material
Infection that occurs in association with bladder catheters, intravascular devices, hip prostheses, or inanimate foreign material, which gains access to the tissues following surgical or traumatic injuries, may fail to respond to antimicrobial chemotherapy. Under these circumstances, total eradication of infection is rarely achieved until the foreign material is removed.
If antimicrobial therapy is to be successful these drugs can only exhibit their full power as ‘magic bullets’ if used intelligently and if full recognition is paid to the need to individualize each course of treatment to both the patient and the pathogen.