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 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. The benefits and risks of chemoprophylaxis depend on:
Prevention strategies for infectious disease can be characterized by the traditional concepts of primary, secondary, and tertiary prevention.Primary prevention can be defined as the prevention of infection. Secondary prevention includes measures for the detection of early infection and effective intervention before symptoms occur. Tertiary prevention consists of measures to reduce or eliminate the long-term impairment and disabilities caused by established infection.
Failure to consider fully the risks of prophylaxis or to be realistic about the benefits has made unnecessary chemoprophylaxis one of the commonest forms of antibiotic misuse. None the less it is important to recognize that the judgement about what is and is not necessary chemoprophylaxis may require complex decisions about the balance of benefits and risks to individual patients and to the population. The example of prevention of neonatal infection by Group B Streptococci (see below) shows how national guidelines committees can produce different recommendations based on their risk assessment of the same evidence.
Most indications for primary chemoprophylaxis involve starting prophylaxis before a period of defined risk (e.g. elective surgery or travel to regions with endemic malaria or intrapartum exposure of neonates to maternal infections) or following exposure to a patient who is known to have a contagious, dangerous infection (e.g. meningococcal meningitis).
Prophylaxis in surgery
Reducing the risk of surgical infection is probably the commonest indication for chemoprophylaxis and accounts for up to a third of total antibiotic use in an acute hospital. There is no doubt that prophylaxis can reduce the risk of surgical infection but at best it is only one component of effective infection control and unnecessary use of prophylaxis puts patients at risk of infection by Clostridium difficile or antimicrobial resistant bacteria with no compensating benefit.
Classification of operations by risk of infection
The aim of chemoprophylaxis is to reduce the risk of surgical site infection, meaning infection in any part of the operative field from the superficial wound down to the deepest tissues involved in the operation. Postoperative infections can occur at other sites, for example the respiratory or urinary tract but chemoprophylaxis is targeted at surgical site infection.
The probability of surgical site infection is determined by the risk of contamination of the wound, the patient's general health, and the duration of the operation. Risk of contamination is defined by classifying wounds as clean, clean contaminated, contaminated, or dirty (Table 17.1). Primary prevention by chemoprophylaxis is only achievable for clean, clean-contaminated, or contaminated wounds because the process of infection has started pre-operatively in dirty wounds. The risk of infection rises progressively with the degree of contamination of the wound; however, the patient's co-morbidities and duration of surgery are equally important determinants (Table 17.2). A prolonged operation with a clean wound in a patient with co-morbidities carries a 5% risk of wound infection, substantially higher than the 3.5% risk for a contaminated wound in a patient with no co-morbidities and a short operation (Table 17.2).
In addition to the probability that an infection will occur it is important to consider the consequences of infection for the patient and the health service. Surgical site infection following colon surgery is associated with substantially increased risk of mortality and prophylaxis significantly reduces death within
30 days of surgery. An increasing proportion of surgical procedures involve the implantation of devices such as prostheses (e.g. artificial joints) or cardiac pacemakers. Postoperative infection can rarely be controlled without removal of these devices, resulting in long-term morbidity, which can be reduced through surgical prophylaxis. Because of these dire consequences even small reductions in the risk of surgical site infection by prophylaxis may be justifiable.
Table 17.1 Classification of surgical operations according to risk of contamination of the wound by bacteria. Reproduced from Scottish Intercollegiate Guidelines Network (SIGN), Guideline 45. Antibiotic Prophylaxis in Surgery, 2000. http://www.sign.ac.uk with permission.
Table 17.2 Probability of wound infection by type of wound and risk index. The risk index includes the American Society of Anesthesiologists risk score based on co-morbidities and a score based on duration of the operation (see text). Reproduced from Scottish Intercollegiate Guidelines Network (SIGN), Guideline 45. Antibiotic Prophylaxis in Surgery, 2000. http://www.sign.ac.uk with permission.
Principles of surgical prophylaxis
The key to successful surgical prophylaxis is to achieve effective antimicrobial concentrations in the wound before bacterial contamination occurs (Fig. 17.1). Prophylaxis can still be effective if it is administered within 2 h after the start of the operation but later administration is much less effective. However, prophylaxis should not be administered too early before the start of the operation or antibiotic concentrations in the surgical site will have declined below effective levels before bacterial contamination occurs. The ideal time to administer intravenous prophylaxis is in the anaesthetic room no more than 30 min before the start of surgery and before any tourniquet is applied to reduce blood supply to the surgical site. For drugs with a short half-life (2 h or less) additional doses may be required during the operation if it is prolonged or if there is substantial blood loss (1500 ml or more) or haemodilution by 15 ml/kg or more.
Fig. 17.1 Risk of wound infection by time of administration of antibiotic prophylaxis. Data from a prospective study of 2847 patients undergoing elective clean or clean-contaminated surgery who received prophylactic antibiotics. The timing of prophylaxis refers to the time of administration of the first dose in relation to the start of the operation. From Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. New England Journal of Medicine 1992; 326: 281-286.
It follows that there is no value in administering additional doses of prophylaxis once the wound has been closed. For most operations a single dose of prophylaxis is all that is required. There is still some controversy about the benefits and risks of extending prophylaxis for up to 24 h postoperatively, the argument in favour being that wounds are insufficiently sealed to prevent bacterial contamination. However, there is absolutely no doubt that prophylaxis should not continue for more than 24 h after surgery.
Guidelines for surgical prophylaxis
Unequivocal evidence for the effectiveness of antibiotic prophylaxis comes from controlled clinical trials with a ‘no treatment’ or placebo arm. Once effectiveness has been established for a specific operation it may be considered unethical to do placebo-controlled trials in similar operations. For example, the recommendation to give prophylaxis before total knee replacement is based on expert opinion that the strong evidence supporting prophylaxis for total hip replacement supports prophylaxis for all total joint replacements. Guidelines should be explicit about the expected standard of care by distinguishing between: operations for which prophylaxis should be the rule; those for which local policy makers or surgeons may identify exceptions; and operations for which prophylaxis should not be given at all (Table 17.3). This judgement is based on a combination of the evidence about effectiveness of prophylaxis and estimates of the consequences of surgical site infection for the patient.
However effective prophylaxis is at reducing the risk of surgical site infection there must be a point when the risk of infection is so low without prophylaxis that the benefits are questionable. The number of patients that must receive prophylaxis to prevent one surgical site infection rises exponentially as the risk of infection diminishes (Fig. 17.2). For clean wound surgery in a patient with no other risk factors the probability of surgical site infection is only 1% (Table 17.2). If prophylaxis halves the risk of infection that means that 200 patients must receive prophylaxis in order to prevent one surgical site infection. The balance between benefits and risks of prophylaxis therefore depends on how a surgical site infection will be managed if it occurs. An infected hip prosthesis may require 2 months or more of antibiotic treatment (over 200 doses of flucloxacillin); consequently, administration of single doses of prophylaxis to 200 patients is still likely to reduce total antibiotic use in the hospital if it prevents one infection.
Table 17.3 Examples of recommendations for surgical prophylaxis in a national guideline. Reproduced from Scottish Intercollegiate Guidelines Network (SIGN), Guideline 45. Antibiotic Prophylaxis in Surgery, 2000. http://www.sign.ac.uk with permission. Examples of recommendations for surgical prophylaxis in a national guideline. Adapted from:Antibiotic Prophylaxis in Surgery. Scottish Intercollegiate Guidelines Network
Fig. 17.2 The relationship between the number of patients who must receive surgical prophylaxis to prevent one surgical site infection and the risk of infection without prophylaxis. The three lines show results for operations in which the odds ratio of infection with prophylaxis versus no prophylaxis is 0.3 (- · -), 0.4 (—) and 0.5 (….).
The choice of agents for chemoprophylaxis should be guided by local sensitivities and antibiotic policies. The only requirement is that the regimen covers the common pathogens that cause infection at the surgical site. The single dose for prophylactic use is in most circumstances the same as would be used therapeutically.
Prophylaxis for travellers
Protecting travellers to areas where malaria is common fulfils all of the criteria for successful chemoprophylaxis. 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. Although resistance is an increasing problem in malaria it is still possible to identify a drug that is suitable for prophylaxis for most travellers (see p. 425-6). Failure to take chemoprophylaxis or to continue for 4 weeks after leaving a region of endemic malaria are the commonest reasons for malaria presenting in countries where malaria is not endemic.
As in surgery, chemoprophylaxis for malaria should be seen as only one component of risk reduction. They should also do all that they can to avoid contact with malaria vectors: use of protective clothing; window netting; insect repellants; and sleeping under insecticide-impregnated bed netting.
Apart from transient visitors to endemic regions the only other clear candidates for antimalarial chemoprophylaxis are resident women during pregnancy.
Other prophylaxis for travellers
Chemoprophylaxis can reduce the risk of traveller's diarrhoea but early treatment of symptomatic cases is highly effective and is preferable.
Primary prevention of other bacterial infections
Prevention of bacterial endocarditis in people with abnormal heart valves or other endocardial disease is similar in principle to surgical prophylaxis. The aim is to achieve effective plasma concentrations of antimicrobial drugs before the start of a procedure that carries substantial risk of causing bacteraemia and therefore contamination of the endocardium. The logic behind routine prophylaxis for dental procedures is that bacteria from the normal oral flora account for a substantial proportion of cases of endocarditis and that recent dental treatment is a known risk factor. Having said that, there is little direct evidence on the reduction in risk achieved by chemoprophylaxis for dental procedures, and even less for endoscopy of the gastrointestinal or urinary tract. None the less, national and local guidelines recommend prophylaxis for patients at risk undergoing a wide range of procedures that may cause bacteraemia. This is one of the situations in medicine where we have to accept that we will never have high-quality evidence because of the ethical and practical difficulties of doing placebo-controlled trials.
Selective decontamination of the digestive tract
Chemoprophylaxis has a role in the prevention of ventilator-associated pneumonia but once again, only as part of a risk reduction strategy that includes other effective measures. Selective decontamination of the digestive tract, started at the time of intubation and targeted at Gram-negative aerobic bacilli and fungi, reduces the risk of ventilator-associated pneumonia, especially in trauma victims.
Another example of selective decontamination of the digestive tract is oral administration of co-trimoxazole or quinolones to afebrile neutropenic patients. However, this is no longer recommended practice in most national guidelines because the risks from selection of drug-resistant bacteria are thought to outweigh any clinical benefit.
Neonatal infection with Group B streptococci
Colonization of the vagina or rectum with Group B streptococci (Streptococcus agalactiae) is very common in pregnancy (prevalence 20-30%). It does not usually cause morbidity in mothers but can cause serious neonatal infections (10% overall mortality but 23% mortality in premature infants of less than 35 weeks' gestation). However, routine intrapartum chemoprophylaxis does not significantly reduce neonatal mortality and most guidelines recommend targeted intrapartum prophylaxis. There are three different approaches, each advocated by at least one set of national guidelines:
The arguments for and against are based on estimation of the benefits, costs, and risks of each strategy. The reason that the UK guidelines do not favour routine bacteriological screening is that they estimate that at least 24000 women would need to be screened and about 700 women receive intrapartum prophylaxis to prevent one neonatal infection, whereas with risk-based assessment the number of women who must be treated to prevent one neonatal infection is about 200. The risk assessments and decisions made by the national guideline groups will have been heavily influenced by the different legal systems in the UK and North America.
Post-exposure prophylaxis of bacterial infections
Examples of primary post-exposure chemoprophylaxis for bacterial infection include:
For N. meningitidis there are two distinct situations. The first is prophylaxis of contacts of sporadic cases, which is confined to close contacts (e.g. household or mouth-kissing contacts and healthcare workers who have been heavily exposed to respiratory droplets or secretions). The second is in epidemic situations, where chemoprophylaxis may be used to supplement vaccination strategies for Group A or C strains and for epidemics caused by Group B strains. Sulphonamides are no longer effective for prophylaxis of meningococcal infection because of drug resistance; the current agents of choice are rifampicin or ciprofloxacin.
Primary prevention of recurrent bacterial infections
All of the previous examples involve short-term risk reduction but there are a few situations where risk reduction has to be continued for prolonged periods, for example prevention of bacterial infection following splenectomy or rheumatic carditis, or prevention of recurrent infections.
For recurrent infections it may be possible to identify repeated short periods of risk. For example, some women with recurrent urinary tract infection may be able to achieve satisfactory control by taking a single dose of antibiotics following sexual intercourse, which is really repeated post-exposure prophylaxis.
Primary prevention of viral infections
An example of prophylaxis before exposure is intrapartum treatment of mothers who are known to be HIV positive in order to reduce the risk of perinatal infection. The principles of prophylaxis here are the same as for neonatal Group B streptococcal infection.
Examples of post-exposure antiviral chemoprophylaxis include exposure to influenza A virus (amantadine, rimantadine), influenza A or B virus (zanamivir or oseltamivir), and HIV (antiretroviral drugs). There are two common indications for post-exposure prophylaxis against HIV: after occupational or environmental exposure and after sexual exposure.
Secondary prevention strategies involve the identification of early or asymptomatic infection with subsequent treatment so that such infections are eradicated and sequelae are prevented. Although most secondary prevention programmes involve intervention at the individual level through the use of chemoprophylaxis, they may also operate within the context of a population-based or institution-based screening effort. Routine screening programmes for sexually transmitted diseases such as Chlamydia infection are examples of secondary prevention strategies. Contact investigations for partners of persons with sexually transmitted diseases are also part of a secondary prevention strategy focused on those with known exposure. Another example of a secondary prevention programme that uses chemoprophylaxis is the screening of high-risk populations for tuberculosis infection and subsequent therapy with an antimicrobial drug such as isoniazid to prevent active disease.
Tertiary prevention efforts are measures to eliminate long-term impairment and disability that may result from an existing condition. Because most infectious diseases are treatable, tertiary prevention activities are less common than those used with chronic diseases such as hypertension, diabetes, and coronary artery disease. However, this concept is still applicable to the control of infectious diseases inasmuch as some viral infections are chronic and cannot be eradicated and there are a number of latent infections that cause symptoms only in immunosuppressed patients.
Treatment of HIV infection now relies on highly active antiretroviral therapy (HAART) with combinations of drugs that reduce viral load to undetectable levels and restore counts of CD4 lymphocytes to >200/mm3. Provided that the patient can tolerate the treatment and there is no major resistance to antiviral drugs in the infecting strain of HIV, this treatment can maintain normal immunity for many years and prevent recurrence of latent opportunistic infections (see Chapter 28). However, once the CD4 lymphocyte count falls below 200/mm3 the cumulative risk for developing an AIDS-defining opportunistic infection is 33% by year 1 and 58% by year 2, so chemoprophylaxis must be considered.
Latent infection means that many patients are already infected with these organisms before they become immunosuppressed. With normal immunity the infections may have been asymptomatic from the start or may have caused a transient illness but then remain dormant for very long periods. Pneumocystis carinii rarely if ever causes symptomatic infection in people with normal immunity, whereas herpes simplex does cause recurrent symptoms even with normal immunity. However, in immunosuppressed people the frequency and consequences of symptomatic infection are greatly increased. Other indications for tertiary prophylaxis of latent infection include:
The general principles of chemoprophylaxis are the same for long-term suppression of latent infection as for short-term primary or secondary prevention. However, with long-term prophylaxis the issues of inducing resistance, risk of adverse effects, and cost-effectiveness are even more important.
Resistance or cross-resistance has become increasingly common with prophylaxis for the Mycobacterium avium complex, fungal infections, and P. carinii. For example, following prophylaxis with clarithromycin for M. avium complex infections over half of infections that occur are caused by strains that are resistant to clarithromycin, delaying clinical recovery and requiring treatment with alternative less effective or more toxic agents. Prophylactic regimens may also lead to the development of cross-resistance against more common pathogens. For example, rifabutin used for prophylaxis of infections with opportunistic mycobacteria may result in the emergence of rifampicin-resistant strains of M. tuberculosis and use of antibiotics, such as clarithromycin, azithromycin, or co-trimoxazole may lead to the development of resistance among organisms such as pneumococci that were not the primary targets of prophylaxis.
Patients with HIV infection are particularly vulnerable to adverse reactions to drugs. The reason is probably a combination of immunomodulation or other HIV-related idiosyncrasy, pre-existing organ damage in late stage disease and interaction between the large number of potentially toxic drugs that these patients require.
The balance of risks and benefits of taking long-term prophylaxis versus treating symptomatic infections when they occur determine cost-effectiveness. The issue is not purely financial: the costs of prophylaxis include the inconvenience to the patient of having to take large numbers of medicines all the time and their negative impact on quality of life. In general long-term prophylaxis is cost-effective for P. carinii or opportunistic mycobacteria but not for fungal infection or cytomegalovirus.
Beltrami EM, Williams IT, Shapiro CN, Chamberland ME. Risk and management of blood-borne infections in health care workers. Clinical Microbiology Reviews 2000; 13: 385-407. Available from: http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf
Clumeck N, de Wit S. Prevention of opportunistic infections in the presence of HIV infection. In: Cohen J, Powderly WG (eds). Infectious Diseases, 2nd edn. Mosby, St Louis 2004, Chapter 123.
Public Health Service Task Force. Interventions to reduce perinatal HIV-1 transmission in the United States. 2005. Available at: http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf
Royal College of Obstetricians and Gynaecologists. Guideline No 36: Prevention of early onset neonatal Group B streptococcal disease. Available at: http://www.rcog.org.uk/resources/Public/pdf/summaryGroupB_strep_no36.pdf
Scottish Intercollegiate Guidelines Network, Guideline 45: Antibiotic prophylaxis in surgery, 2000. Available at: http://www.sign.ac.uk