Background

Antimicrobial resistance arises via the selection, during exposure to antibacterial drugs, of inherently resistant species, or through the emergence of resistant variants within hitherto sensitive species.5 Resistance in previously susceptible species can evolve by mutation, and be passed vertically within the species, or it can result from the horizontal acquisition of genetic material from other bacteria. Genetic material is most often transferred as plasmids (self-replicating loops of DNA) or transposons ('jumping genes' - discrete segments of DNA, capable of transfer from one plasmid to another, or to a chromosome).5 Direct uptake of DNA (transformation) can occur in a few species such as pneumococci. Resistances may thus pass between species, including from commensals to pathogens and vice versa. Multiresistance can result from the acquisition of plasmids encoding multiple resistance mechanisms, from the activation of endogenous multidrug efflux systems, or from the cumulative acquisition of resistances through several separate mechanisms.

Excessive and inappropriate use of antibiotics is believed to be one of the most important factors in increasing the prevalence of antibiotic resistance.1^,2^,5^,6 In humans, 80% of antibiotic prescribing occurs in the community. In hospitals, the volume of prescribing (usually initiated by junior doctors) is less, but the concentration of antibiotic use greater; about 20-30% of this is for prophylaxis against surgical infection.1 About half of all antibiotic use in the UK and the USA is in animals, much of it for mass treatment or prophylaxis and for growth promotion.3^,6 A large proportion of all human and animal antibiotic use is considered to be of questionable value.3^,6 Overuse of broad-spectrum antibiotics has particularly been blamed for increasing the prevalence of certain resistances. For instance, enterococci, which are inherently resistant to cephalosporins and fluoroquinolones, have increasingly emerged as pathogens as the use of these drugs has increased.7–9 Exposure to cephalosporins is also associated with rapid acquisition of methicillin-resistant Staphylococcus aureus (MRSA) by hospital patients previously carrying methicillin-susceptible strains.10 Poor adherence to antibiotic prescribing guidelines aggravates the problem of overuse,1 while the availability of over-the-counter antibiotics in many parts of the developing world encourages widespread, indiscriminate and inappropriate use.11

Once selected in one patient, resistant bacteria may spread to others. Many believe that spread is facilitated by high-pressure healthcare systems:1 rapid bed turnover, frequent transfers between wards, hospitals or nursing homes, overcrowding, and overstretching of medical and nursing staff all have the potential to undermine basic measures for infection control.12 Moreover, patients discharged from hospital care can carry resistant organisms into the community, with the result that nursing and residential homes provide large reservoirs for potential spread of resistance and its subsequent reintroduction into hospitals.

Resistant and multiresistant organisms are particularly important as a cause of hospital-acquired infection, especially in immunocompromised, debilitated or elderly patients, for whom they can pose a serious threat to life. At one London teaching hospital around 6-7% of patients contracted an infection during their stay.1 Hospital-acquired infections are often difficult and expensive to treat. They delay discharge and sometimes demand isolation procedures, which further increases the cost of therapy. Within hospitals, multiresistance is most prevalent where antimicrobial use is greatest, notably in intensive care units (ICUs) and other high-dependency units.

Multiresistant pneumococci and Mycobacterium tuberculosis have already become a cause of serious infection in the community in both the industrialised and the developing world. Other pathogens, such as multiresistant Salmonella typhi, cause many deaths in countries where they are endemic, and may be spread by increasing international travel.

We now consider some of the organisms where multiresistance is of greatest significance or potential significance.

Methicillin-resistant Staphylococcus aureus (MRSA)

MRSA in hospitals

MRSA has been a well-recognised cause of hospital-acquired infections worldwide since it was first detected in Europe in the early 1960s.13 The organism can survive for long periods in both the hospital and the home environment and can colonise the skin, nose or throat of patients and healthcare staff and the environment. It is readily spread by direct contact, usually via the hands. MRSA rarely invades intact skin, but can invade pressure sores, surgical wounds, intravascular and bladder catheter sites, and may then give rise to severe infections (e.g. wound infection, bacteraemia, endocarditis, osteomyelitis).14 In England and Wales, the prevalence of MRSA has steadily risen - from 1-2% of reported blood and CSF S. aureus isolates in 1989-9215 to 31.7% in 1997.2 This rise reflects the increasing domination by two 'epidemic' strains (EMRSA-15 and EMRSA-16). MRSA remains uncommon in countries that have strictly applied infection control policies, notably the Netherlands and Scandinavia.16

Treatment of MRSA

Many MRSA remain susceptible to combination therapy with fusidic acid plus rifampicin, or to therapy with an aminoglycoside, a fluoroquinolone, a macrolide or trimethoprim. These drugs are suitable for treating susceptible infections that are not severe or life-threatening. However, mutational resistance emerges particularly readily to fluoroquinolones, fusidic acid and rifampicin and empirical therapy with a single drug is, in general, best avoided. Some MRSA strains are resistant to all antimicrobials except the glycopeptides, vancomycin and teicoplanin. In practice, any severe suspected MRSA infection requires urgent treatment with a glycopeptide. Sometimes only vancomycin is active, so it is extremely disturbing that S. aureus with intermediate-level resistance to vancomycin ('VISA') (minimum inhibitory concentration [MIC]: vancomycin 8mg/L) has already been seen in clinical isolates of MRSA in Japan, the USA and France.17 VISA is difficult to detect by conventional laboratory procedures and has been associated with therapeutic failure. It tends to emerge in patients who have received multiple courses of vancomycin. It is very important that vancomycin is used only when appropriate and essential, both to prevent the spread of VISA strains, and to lessen the selection pressure for the eventual emergence of MRSA with transferable, high-level glycopeptide resistance.17–19

Preventing spread of MRSA

In hospitals, scrupulously applied infection control policies, co-ordinated and monitored by an adequate infection control team, reduce the risk of MRSA infection and the spread of epidemic strains, improve patient care and, on the available evidence, are cost-effective.12^,14^,20 Control is founded on sound antibiotic policies and practices, effective alert systems, and strict compliance with basic control measures (handwashing, aseptic techniques, ward cleaning, handling of waste, use of disposable gloves and aprons, and, where appropriate, isolation or 'cohort nursing' of patients colonised with MRSA).20

Infected or colonised patients form the main reservoir of MRSA in hospitals. Healthcare workers are often transiently colonised during outbreaks, but are seldom the source of an outbreak.21 A combined working party of the British Society of Antimicrobial Chemotherapy, the Hospital Infection Society and the Infection Control Nurses Association has recently updated its guidelines on infection control for MRSA in hospitals.20 The guidelines cover a range of inpatient settings, from high-risk areas (such as ICUs, special care baby units, cardiothoracic and transplant units) to minimal-risk areas (such as psychiatry and elderly long-stay units), in hospitals with or without established MRSA. The report includes recommendations for screening and management of MRSA-colonised patients and staff in these settings. These deserve to be widely studied.

MRSA in the community

Discharged patients colonised with MRSA may introduce the organism into nursing homes where it may be difficult to prevent cross-colonisation. UK-based studies are few, but in one study of 10 nursing homes in the West Midlands, MRSA was isolated from 33 of 191 (17%) elderly residents.22 However, only one resident had a clinical infection. Hospital admission within the previous year, especially for surgery, significantly increased the likelihood of MRSA carriage.

MRSA in the community poses little or no threat to family and friends or to staff or healthy residents in nursing homes. It is more of a hazard in nursing homes that admit high-dependency or early postoperative patients, who may be more susceptible to infection. Moreover, nursing homes are a potential reservoir for hospital outbreaks when colonised residents are readmitted to acute hospitals.

No special precautions are needed for most MRSA-colonised patients discharged into the community:23 colonisation with MRSA should not be a bar to patients going into a nursing home, nor to normal social mixing. Basic precautions of handwashing, cleaning of equipment and handling of potentially contaminated dressings, catheters and linen should be applied routinely, and any wounds and pressure sores on residents, or skin lesions on staff or residents, should be protected by an impermeable dressing.23 Hospitals and nursing homes must communicate effectively, and the role of community infection control nurses needs to be developed, both to teach and implement infection control.23 The advice of the consultant in communicable disease control should be sought if the number of infections in a home rises, or if there is frequent traffic of MRSA-colonised patients between the nursing home and the hospital. In homes with high-dependency patients, control measures should resemble those in hospitals.

Please click here to view tables showing how to treat MRSA infections and carriers.

Enterococci

Enterococci are increasingly frequent pathogens in hospitals. They are found in the stools of healthy people and can cause endogenous urinary tract and wound infections. More seriously, enterococci are the causative organisms in 5-15% of patients with bacterial endocarditis24 and may cause bacteraemia, meningitis and other invasive infections in immunocompromised patients, especially those on transplant, dialysis and oncology units. Enterococcus faecalis is the most frequent species cultured from clinical specimens, but Enterococcus faecium has greater inherent resistance, notably to penicillins.7

Conventional treatment of serious enterococcal infections has involved the use of synergistic combinations of an aminoglycoside with a beta-lactam or a glycopeptide. However, this approach is now being weakened by high-level aminoglycoside resistance, which is found in 26% of isolates from patients with enterococcal endocarditis,24 with concomitant loss of synergism.7^,8 Beta-lactams alone are ineffective against most strains of E. faecium,7 and resistance to glycopeptides, already a problem in the USA, is increasingly reported from hospitals in the UK, again mostly among E. faecium.7^,8 Enterococci resistant to all synergistic combinations are now being encountered, and a few isolates are resistant to all antibacterial drugs.

The emergence of glycopeptide resistance in enterococci is especially worrying. The resistance genes are often readily transferable between enterococci, and their transfer to S. aureus has already been demonstrated in vitro, though not yet in nature.25 Glycopeptide-resistant enterococci have been isolated from sewage and farm animals.8 Use of another glycopeptide, avoparcin, as an agricultural food additive (now banned in the EU), has been implicated in the emergence of such strains.1^,8 It is postulated that glycopeptide-resistant entercocci may reach humans via the food chain, although the frequency of such transmission is controversial.1^,8

Within hospitals, selection pressure for glycopeptide resistance comes especially from use of vancomycin itself (e.g. in treating MRSA infections) and from the widespread administration of cephalosporins, to which enterococci are inherently resistant.7^,8^,19 Once established in a hospital, glycopeptide-resistant enterococci often spread from patient to patient via the hands of hospital staff and via environmental contamination.7^,8^,19 Evidence is needed on optimum management of individuals colonised with a resistant strain.8 Strict barrier nursing, combined with surveillance cultures for patients in high-risk areas and restricted use of vancomycin, have successfully eradicated localised hospital outbreaks of glycopeptide-resistant enterococci.9 In some hospitals, a change in the antibiotic prescribing policy, with use of penicillin/beta-lactamase inhibitor combinations instead of cephalosporins, has substantially reduced infection and colonisation with vancomycin-resistant enterococci where the above control measures have failed.9

Streptococcus pneumoniae

The development and rapid spread worldwide of penicillin-resistant pneumococci is a major concern. Pneumococcus is the foremost cause of community-acquired pneumonia and otitis media, and a major cause of bacterial meningitis. High-level penicillin resistance (MIC 2mg/L or greater) was first reported from South Africa in the 1970s, but by 1992-93 was found in 23% of lower respiratory isolates from 14 centres in Western Europe and the USA, and in around 50% of those from Spain and France.26 In England and Wales, the prevalence of intermediate-level (MIC 0.1-1mg/L) and high-level penicillin resistance combined was only 3.9% in March 1995, but this was still more than double that found 5 years earlier.27 Moreover, the prevalence of resistance varies markedly with locality, being up to 20% in some regions,28 emphasising the importance of good local surveillance of resistance.

The prevalence of erythromycin-resistant pneumococci also increased in England and Wales, from 2.8% in 1990 to 8.6% in 1995.27 Penicillin-resistant pneumococci are more likely than penicillin-susceptible strains to be multiresistant to unrelated antibiotics, including tetracyclines, co-trimoxazole and chloramphenicol as well as macrolides.26^,28

Treatment in the community

At present, oral amoxycillin (0.5-1g every 8 hours in adults; 25-50mg/kg/day in three divided doses in children) remains a good first choice for empirical treatment of patients with suspected pneumococcal pneumonia who are well enough to be managed in the community.29^,30 A macrolide is usually a suitable alternative in those allergic to penicillin. Amoxycillin (paediatric dose 125mg three times daily, or 750mg twice daily for 2 days) is also probably still the best choice for empirical therapy of otitis media, even in areas where penicillin resistance is prevalent among pneumococci.29 If combined with a beta-lactamase inhibitor, amoxycillin also covers other common middle ear pathogens, notably Haemophilus influenzae and Moraxella catarrhalis. Oral cephalosporins are consistently less active than amoxycillin against pneumococci, particularly strains with intermediate-level penicillin resistance.29

Treatment of patients in hospital

In more severely ill, hospitalised patients with pneumonia or bacteraemia, penicillin given in adequate doses (ampicillin 1g i.v. every 6 hours, or benzylpenicillin 1.2g i.v. every 4-6 hours) remains effective against pneumococci with intermediate or even high levels of resistance, unless the patient is immunosuppressed.29–31 The position is less clear for strains with very high levels of penicillin resistance (MIC above 4mg/L).30^,31 Treatment should be guided by culture results, frequent monitoring of the patient's condition and the advice of the consultant microbiologist or infectious diseases specialist.

Penicillin and ampicillin are likely to be ineffective against meningitis caused by pneumococci with intermediate- and high-level resistance to penicillin, and are unsuitable for empirical therapy of this life-threatening disease.29 Cefotaxime or ceftriaxone are the preferred drugs for initial empirical treatment of bacterial meningitis outside the neonatal period, but resistance has also been reported with these drugs among pneumococci, leading to treatment failure.29 Consequently, some experts now recommend addition of vancomycin until the antibiotic susceptibility of the infecting organism is known.29^,32

Gram-negative bacteria

Hospital-acquired infections

Resistant and particularly multiresistant Gram-negative bacteria are increasingly important causes of hospital-acquired infection, especially in debilitated patients with underlying disease or immunosuppression. Numerically the organisms of greatest concern are bowel commensals of the Enterobacteriaceae family, and Pseudomonas aeruginosa, which is ubiquitous in hospitals. Among the Enterobacteriaceae, Escherichia coli, Proteus and Klebsiella spp. can cause endogenous infections in healthy subjects (most often in wounds or the urinary or respiratory tract), and opportunistic infection in compromised patients; Enterobacter and Citrobacter spp. are almost entirely associated with hospital-acquired infections in debilitated or immunocompromised patients, including neonates. A further Gram-negative genus, Acinetobacter spp., is increasingly encountered as a cause of serious opportunistic infections in compromised patients, notably ICU patients requiring artificial ventilation.33 A worrying feature of this last organism is its capacity rapidly to develop multiresistance, including, to some extent, resistance to carbapenems.33

Enterobacteriaceae produce a variety of natural beta-lactamases giving drug resistance.34 Many newer beta-lactams were designed to resist the action of bacterial beta-lactamases. However, Gram-negative bacteria have evolved that possess more potent enzymes, able to break down the new beta-lactams. Some of these enzymes are mutants of older beta-lactamase types and are called 'extended-spectrum beta-lactamases' (ESBLs). The genes encoding ESBLs are often located on transferable plasmids that can pass between several species of enterobacteria, and these plasmids often encode multiple other resistances. Around 25% of Escherichia coli isolates from blood and CSF are now multiresistant (to four or more antibiotics),35 while 20-40% of enterobacter and citrobacter isolates are resistant to all beta-lactams except carbapenems (imipenem and meropenem) and temocillin.34 ESBL-producing multiresistant strains are common among Klebsiella spp. isolated from ICUs in Europe,36 and ESBL-producing Klebsiella pneumoniae have been responsible for major outbreaks of infection in UK hospitals.37 All ESBL-producers should be regarded as resistant to ceftazidime, ceftriaxone and aztreonam, and they are often cross-resistant to aminoglycosides and ciprofloxacin.36

Most clinically important Gram-negative bacteria have so far remained susceptible to the carbapenems. An exception is P. aeruginosa where low-level, narrow-spectrum resistance to carbapenems arises readily by mutations that reduce bacterial cell wall permeability to the drugs.34^,38 The emergence and dissemination, in Japan, of a plasmid-mediated carbapenemase (i.e. a carbapenem-destroying enzyme conferring high-level resistance to all beta-lactams including carbapenems) is a more worrying development.39 A different carbapenemase, giving similar resistance, has recently been found in a UK isolate of P. aeruginosa.38

Community-acquired infections

In the community, the development of resistance to amoxycillin, now present in about 50% of urinary pathogens, has made the drug unsuitable for empirical treatment of urinary tract infection.40 Trimethoprim remains effective against around 70% of urinary pathogens; nitrofurantoin (resistance rate 15%) or co-amoxiclav (resistance rates around 10%) are useful alternatives.40 A 3-day course is recommended for otherwise healthy women with uncomplicated cystitis.2^,40 Microbiological investigation and sensitivity testing are needed in women who are pregnant and in patients with suspected pyelonephritis, or recurrent or complicated infections related to underlying structural abnormalities, recent instrumentation or catheterisation. To preserve their efficacy, fluoroquinolones should be reserved for treating patients with complicated urinary tract infection or infection with an organism resistant to other antibiotics.40

Salmonella infections are common in the community. The development of multiresistance in some strains has meant that, in the event of invasive disease (e.g. bacteraemia, meningitis), treatment may be difficult. Salmonella typhimurium DT104 is one of the most common isolates in humans in the UK.41 A multiresistant phenotype of DT104, with chromosomally encoded resistance to ampicillin, chloramphenicol, streptomycin, sulphonamides and tetracycline, has caused many household or community outbreaks, and often causes severe illness requiring hospital admission. Many isolates of type DT104 now show additional resistance to trimethoprim.41 Such resistances have made ciprofloxacin the preferred drug for patients with invasive salmonella infections. However, around 12% of blood-culture isolates in England and Wales now have reduced susceptibility to ciprofloxacin too, and this has sometimes been associated with treatment failure.41

In areas of the world where typhoid fever is endemic, multiresistance of Salmonella typhi to chloramphenicol, ampicillin, trimethoprim, streptomycin and sulphonamides, has become widespread.42 In the UK, which has 200-300 cases of typhoid fever annually (mostly in people returning from the Indian subcontinent), the proportion of chloramphenicol-resistant isolates rose from around 1% in 1989 to 35% in 1994, and most of these were also resistant to ampicillin and trimethoprim.42 Again, ciprofloxacin is currently the drug of first choice. Worryingly, however, reports of resistance in S. typhi are also beginning to appear:42 indeed, multiresistant strains with resistance to ciprofloxacin have already caused major outbreaks of typhoid fever in Asia.

Multiresistance to ampicillin, tetracycline, sulphonamides and trimethoprim has also markedly increased in Shigella spp. in many parts of the developing world, where the organism is a major cause of acute invasive and often fatal diarrhoeal disease.11 In England and Wales over 70% of isolates of Sh. dysenteriae, Sh. flexneri and Sh. boydii (organisms mainly acquired during foreign travel), and 45% of Sh. sonnei (which is indigenous in the UK) are resistant to four or more antibiotics.43 Most UK isolates remain susceptible to nalidixic acid (the choice in children when empirical treatment is needed) and ciprofloxacin or olfloxacin (the choice in adults).43

The recently reported finding, in Vietnam and France, of transmissible high-level resistance to chloramphenicol, sulphonamides and streptomycin in strains of Neisseria meningitidis type B44 is also worrying, especially for the developing world, where chloramphenicol is cheap and widely used to treat meningitis. Most strains of N. meningitidis remain fully sensitive to penicillin and other beta-lactams. Relative insensitivity to penicillin, reported from many countries including the UK, can be overcome by treatment with penicillin in high doses.45 Beta-lactamase-producing (and, therefore, fully penicillin-resistant) strains are still rare.45 Finally, the emergence of fluoroquinolone resistance in Neisseria gonorrhoeae, already frequently resistant to sulphonamides, penicillins and tetracycline, means that, in some areas (e.g. parts of the Far East), only third-generation cephalosporins are reliably effective for the treatment for gonorrhoea.11^,46 In the UK, fluoroquinolone resistance in N. gonorrhoeae so far remains uncommon.46

Mycobacterium tuberculosis

Multidrug resistance in Mycobacterium tuberculosis (MDRTB), defined as resistance to at least isoniazid and rifampicin, is now a worldwide problem. Between 1994 and 1997, the Global Surveillance Project of the World Health Organization and the International Union against Tuberculosis and Lung Disease identified multiresistant strains in all 35 participating countries, from five continents, with a median prevalence of 2.2% of isolates (range 0-22%).47 MDRTB were especially prevalent in the Baltic States, Russia, India, Argentina and the Dominican Republic. In England and Wales, the prevalence of MDRTB in initial isolates rose from 0.6% to 1.7% between 1993 and 1996; in London, 1.9% of isolates were multidrug resistant.48

Primary drug-resistant tuberculosis is the result of transmission of a resistant strain to someone with no previous history of the disease. Acquired resistance, found in people previously exposed to treatment, results from inadequate therapy, for instance, because of an inappropriate regimen (e.g. a single drug, or treatment of inadequate duration), non-adherence to treatment, or interruption to the supply of medication.47 Infection with MDRTB is associated with high mortality, particularly in immunosuppressed or malnourished patients. It is also associated with prolonged sputum positivity and therefore a higher risk of transmission to close contacts including family, healthcare workers and other patients. Cross-infection with MDRTB is more likely in institutions where high-risk patients (e.g. people infected with HIV) are gathered, such as hospitals, hostels and prisons. In these situations, special attention should be paid to rapid diagnosis of MDRTB and isolation, in a negative pressure ventilated room, of sputum-positive patients.48–50

'Directly observed treatment, short course' (DOTS) may have an important role in helping to ensure that patients with tuberculosis complete an adequate course of therapy, thus minimising the selection of acquired resistance.49 However, short-course therapy based on isoniazid, rifampicin and pyrazinamide will not cure patients already ill with MDRTB. Such patients need lengthy and fully supervised treatment with several drugs, with close monitoring, both for non-adherence to the treatment and for toxicity.49 Nor is DOTS appropriate in areas where MDRTB is well established; there, therapy has to be tailored to the susceptibility of the organism and, again, often involves prolonged courses of second-line drugs.

Responding to resistance

New developments in treatment

In the past, development of new antibiotics has kept pace with the development of resistance. However, new drugs, in turn, often selected for organisms with new patterns of resistance, such as the ESBL-producing Enterobacteriaceae, which can inactivate the new-generation cephalosporins.

Among antimicrobials under clinical development, several show promising activity against infections caused by Gram-positive bacteria.51 Quinupristin/dalfopristin is a synergistic combination of two streptogramins; it is not yet licensed in the UK but has been used on a named-patient basis for the treatment of severe infections with multiresistant E. faecium where other treatments have failed. It is also active against streptococci and staphylococci (including MRSA). Everninomycin, an analogue of a drug previously used in animals, and the oxazolidinones, a new class of antimicrobial drug, are also active against enterococci, pneumococci and staphylococci including MRSA. Several new fluoroquinolones, including grepafloxacin, trovafloxacin and moxifloxacin, have improved anti-pneumococcal activity compared with the earlier fluoroquinolones, while retaining anti-Gram-negative activity against ciprofloxacin-sensitive strains. Disturbingly, however, there is no new class of drug with activity against Gram-negative infections in near prospect.

Reforms at the clinical and strategic level

There is wide agreement that confronting the spread of antibiotic resistance and multiresistance requires more than the development of new drugs. Major efforts are needed to change the public's perception and doctors' use of antimicrobial drugs. Effective strategies to conserve present antimicrobials, to improve infection control and to achieve surveillance of resistance, will need to be developed and resourced. They will require co-operation at local, national and international levels. The recent reports of the House of Lords Select Committee and the Standing Medical Advisory Committee have identified many areas where action is required.1^,2 The recommendations include:

? Greater emphasis on infectious diseases and antimicrobial therapy in undergraduate and postgraduate education; increased input by health authorities towards the professional development of doctors in the field of antibiotic prescribing, especially through prescribing audit and feedback and educational outreach.

? Public education by Government and health authorities and more teaching of communication skills for GPs. These are needed to encourage better understanding about appropriate antibiotic use and to deter inappropriate demand for antibiotics for self-limiting infections such as simple coughs, colds and sore throats.

? The development locally of antibiotic formularies and evidence-based clinical guidelines, reinforced through education and audit.

? Improved documentation of antimicrobial use in hospitals, including computer systems for patient-specific prescribing data at ward level.

? Renewed resolve by purchasers and commissioning agencies, backed by resources, to put infection control and hygiene measures at the heart of hospital management and practice. National standards and guidelines for community infection control are proposed, including the provision of at least one infection control nurse in every district health authority.

? Surveillance for antibiotic resistance patterns to be improved, and put on a comprehensive and strategic footing, in order to inform and support local formularies and antimicrobial prescribing policies, track the increase in resistance, and identify areas of weakness in infection control policies.

The House of Lords Select Committee report draws attention to the need for continued research to determine optimal policies for antibiotic usage, both in treatment and prophylaxis, especially with respect to the duration of antibiotic courses. It highlights microbiological surveillance as a legitimate area for priority support by the NHS Research and Development Directorate and other bodies. Other recommendations concern the drug licensing authorities, the pharmaceutical industry, and the need to encourage, urgently, more prudent use of antimicrobial drugs in veterinary medicine and farming. The recommendations in the report have been widely welcomed.

Conclusion

Resistance and multiresistance to antimicrobial drugs are becoming more prevalent in many common pathogens, and pose serious problems both in hospitals and the community. In several areas, the development of new drugs is no longer keeping pace with accumulating resistance. Limiting the proliferation of resistance will require sustained efforts from governments, healthcare professionals, the pharmaceutical industry and the public. The priorities include: improved professional and public awareness of appropriate antibiotic usage through education; the development and implementation of evidence-based policies for antimicrobial therapy, supported by national and local prescribing guidelines and audit; reliable and adequately resourced policies for infection control in hospitals and the community; systematic surveillance, supported by information technology, of resistance patterns at local, national and international levels, with linkage to antibiotic prescribing data.

Guidelines already exist for many common clinical situations, though not all of these are evidence-based, nor designed specifically to mitigate resistance. Prescribers need to be familiar with recommendations for first-line antibiotic therapy, and with local resistance patterns. Use, especially empirical use, of second-line and newly marketed antibiotics is to be discouraged except in defined clinical situations where there is clear agreement at local or national level. In many situations, the infection will be self-limiting and no antibiotic therapy is appropriate.