## Abstract

Antibiotics have undoubtedly made a major contribution to improvements in both human and animal health and welfare. Recent years have brought alarming rises in the prevalence of resistance to some agents among certain groups of bacteria. Concern is growing that therapeutic options will become increasingly limited if resistance rates continue to rise. There is widespread agreement that action is required to reverse or at least slow this process. Necessary steps to manage the situation include better surveillance to assess accurately the extent of problems, more prudent use of the available antibiotics to conserve valuable therapeutic resources and improved infection control to limit the spread of resistant organisms. Achieving these goals will not be possible without government, the medical profession and the public being better informed and educated. Regulatory bodies and the pharmaceutical industry need to work together to ensure a steady supply of new antimicrobials. Our understanding of the processes driving resistance at both the molecular and population levels is advancing. However, the relative contributions of the various uses of antimicrobials to the resistance problem and which will be the most effective containment measures are still hotly debated. Progress is being made, but continued concerted action is necessary if the usefulness of this most important group of therapeutic agents is to be preserved.

## Introduction

Although mankind flourished over past centuries without antimicrobials, few would disagree with the view that of all the therapeutic agents, antibiotics have had a most important impact on human and animal health. Not only have the major bacterial infections been controlled but far more technologically advanced medicine has been practised with the aid of antibiotics. Food animals can now be reared more intensively (hence cheaply) with the use of growth promoters and herd/flock therapy. However, both advanced technology and modern animal rearing carry major risks to public health.

Antimicrobial resistance has been recognized since the earliest days of chemotherapy. It now involves almost all the genera of bacteria associated with disease in animals and man. The problems appear to be accelerating, accumulating and global. Not only are antibacterials involved but also antifungals, antivirals and antiparasitic agents.1,2 The impact of resistance on animal health is often little appreciated by the medical profession, for example, many species of nematodes of sheep and horses are resistant to anthelmintics: resistance of the common sheep liver fluke to benzimidazoles and avermectins has seriously curtailed sheep farming in Australia, and resistance in Parascaris equorum (an ascarid of horses) has been reported in the Netherlands.3

It would therefore seem that the profligacy of man is endangering the vital resource that antimicrobials bring mankind. What are the risks to man and what can be done to stabilize or even reverse this situation? These are major issues of public health and the answers are far from clear. This review will attempt to address some of these concerns, concentrating on antibacterials.

## The bacteria

To bacteria, antibacterials are, by definition, a threat. So much of our knowledge to date can be very simply explained according to Darwinian principles, popularly described as ‘survival of the fittest’.4 Jones, in his update of The Origin of Species5 points to three properties which favour natural selection in any population. Firstly, a large number of individuals, a condition undoubtedly true for bacteria, which so vastly outnumber all other life forms combined. Jones has calculated that man is exposed to bacteria in excess of 1024 and this excludes those in soil and sea. Secondly, bacteria exhibit a variety of mechanisms to alter or exchange hereditary information, including mutation and transformation and, in particular, conjugation which can be mediated rapidly and efficiently by plasmids and transposons. Additionally, the observation of hypermutation and its contribution to the acquisition of antimicrobial resistance is yet another fascinating insight into the ability of bacteria to adapt to new environments.6 Mankind on the other hand has more limited and slower abilities to adapt to changing environments. Thirdly, natural selection is favoured by the intensity of selection pressure and exposure to any, in this context, antimicrobial use. It is extremely difficult to judge the total global antimicrobial production but it is probably 100–200 million tonnes per annum at a conservative estimate. Use in the less economically developed countries is extraordinarily difficult to assess, as is global animal use. Mankind has overcome many of the selection pressures (war, famine, disease, etc.) yet ironically, of course, antimicrobial resistance may yet impose a new selection pressure.

Thus almost all bacteria seem to express resistance to antibacterials and different environments generate particular problems. Until recently, it was considered that the hospital, and especially the intensive care units (ICUs), were the major source of bacterial resistance, with many reporting problems worldwide with Pseudomonas aeruginosa, Acinetobacter baumanii, the Enterobacteriaceae and staphylococci. The reasons for this are, of course, well known. Antibacterials are used intensively and opportunist bacteria and cross-infection are all too common.

I believe that the effects of community resistance are already evident. Examples include the continuing problems of penicillin and macrolide-resistant Streptococcus pneumoniae,7 now compounded by fluoroquinolone resistance in this organism.8 This profoundly alters the therapeutic approach to community-acquired pneumonia and meningitis in some countries. Whereas methicillin-resistant Staphylococcus aureus (MRSA) was considered a hospital pathogen, it is now increasingly recognized as a community problem, whether this be the hospital endemic strains being found in vulnerable patients such as the elderly in care homes,9 or the more recently described cases with the Panton-Valentine leucocidin virulence gene, which are community-acquired.10 Another example of profound change is the expanding problem of extended spectrum β-lactamases (ESBLs) found in the community for some time in France and more recently in UK hospitals.11 Perhaps ESBL-mediated resistance will become the ‘norm’ just as TEM-mediated resistance to ampicillin is amongst many Gram-negative genera. The increasing resistance of Neisseria gonorrhoeae to the fluoroquinolones is yet another example12 and is having a major impact on therapy.

Why are we seeing this change from hospital to community? Again a possible explanation lies in Darwinian principles. The UK House of Lords Report13 estimated that about 90–95% of all antimicrobial use in the UK is within the community, as opposed to hospitals. In addition, changes in society such as day care centres for children, homes for the elderly, and communities of other groups such as drug abusers, all create fertile grounds for microbial transmission.

## Solutions to the problems?

The tools to manage the problems of antibacterial resistance are remarkably few. As discussed below, novel agents active against new bacterial targets (as opposed to modifications of existing antibacterials) are likely to be few in the foreseeable future. The three remaining interventions, reducing selection pressure (that is reducing antibiotic prescribing), cross-infection control, and the increased use of vaccines all need to be assessed and monitored.

### Surveillance

Although the best definition of surveillance is information for action, very few if any measures of bacterial resistance provide such data. The vast majority of data is passively acquired, that is, it is derived from samples received in hospital laboratories, with little knowledge of why it was sent in the first place and how representative any bacteria isolated are of such strains in the general population. Hence, the main weakness is lack of meaningful denominator data. Without this, such surveillance could be positively misleading. For example, do the Escherichia coli reported from urine over- or under-estimate the resistance in E. coli at large? There is some suggestion that such data overestimate the problem, in this example.14 Without such knowledge, it is difficult to devise realistic guidance for clinicians.

The Seventh Report of the House of Lords Science and Technology Committee13 stressed the importance of surveillance yet also reflects on its piece-meal development in the UK (and elsewhere). Changes are taking place and in the UK mandatory reporting of MRSA bacteraemia and vancomycin-resistant enterococci are an important development but still shed little light on the ‘drivers’ of resistance. There are a number of important issues which need to be addressed.To achieve the above is going to take time and investment. A first step is to undertake this by setting up sentinel studies in the community, and in hospitals, where the ‘antibiotic pharmacist’ can play a central co-ordinating and information-collecting role.

• Laboratory: these include (i) accurate identification of strains. For example, identifying only to the level of ‘coliform’ should be abandoned as far as possible; (ii) often too few antimicrobials are tested; (iii) consistency of methodology is required, and is being addressed by the BSAC Working Party;15 (iv) molecular characterization of resistance will give greater insights.

• Hospital surveillance: (i) improved information on hospital prescribing is central to our understanding of resistance in this setting; (ii) improved information technology (IT); (iii) strategies for monitoring must be co-ordinated with infection control/antibiotic prescribing control measures. The development of ‘antibiotic pharmacists’ in England is an important step in achieving this.

• Community surveillance: (i) attempting to overcome the bias in samples sent to laboratories; (ii) lack of population denominators; (iii) the need for primary care prescribing data; (iv) the need to understand where resistance is initiated—in hospital with spread to community or vice versa; (v) appropriate IT which can report on prescribing, resistance and clinical outcomes.

• Information technology: central to all the above is the need to achieve the following linked data-sets: hospital prescribing—by clinical indication; hospital resistance—by site and organisms and similarly in the community.

The lessons we have failed to learn in antibacterial use and resistance should not be repeated in antiviral therapy. There are now about 30 antiviral drugs available and resistance is an issue, especially in antiretroviral therapy.16 Again there is a major need to have denominator studies of clinical failure rates, the prevalence and nature of antiviral resistance, standardization of laboratory methodology and national (or at least sentinel) databases of the above. The pace of increase and extent of anti-retroviral resistance in sub-Saharan Africa is so alarming that if the investment and hopes in this therapy are not to come to nought, then such surveillance is an immediate priority.

Most developed countries do undertake antibacterial resistance surveillance but, by and large it is, as indicated, of a poor calibre. At best it sheds little light on the problems and at worst it can be misleading. In an excellent example of the latter, Harbarth and colleagues17 have shown the difference between the use of aggregated data (on imipenem use and resistance) where there was little or no relationship between the number of defined daily doses used and susceptibility data, yet when individual cases were analysed, the risk of imipenem resistance in an individual had a marked relationship to exposure to the antimicrobial.

### Infection control

It is not the purpose of this review to examine infection control measures that can reduce the spread of resistant bacteria although it is axiomatic that there is a major role for improvements in this area and the impact it will have. In hospitals, certainly in circumscribed units, such as intensive care units (ITU), enhancement of infection control procedures can have a major impact. Both in the ITU and more widely in a hospital, the interplay between antimicrobial use and the spread of infection is extremely complicated. It is stated that hospital-acquired infection costs the UK £1 billion per year and leads to some 5000 deaths18 yet there are poor provisions in hospitals to contain the problem (for example, sufficient single rooms), too many patient movements from one ward to another, a lack of trained staff and unclear lines of responsibility.

As already mentioned, there is a trend for multi-resistant pathogens, such as MRSA and ESBLs to be spreading in the community. There is a major need to understand the drivers of this new phenomenon so that concerted and informed action can occur.

Further information is required on the impact of two modern developments in society. As mentioned above, at both extremes of life the traditional role of care is changing. The very young are now coming together in child-care centres and the elderly in care homes. The role of both of these settings in the spread of resistant bacteria must be considerable. Far more research is required to understand these issues. To change attitudes will be very difficult, but, for example, requiring children who are receiving antimicrobials not to attend such centres would have a major socio-economic impact. The greater role of immunization of both groups is promising19 and needs further exploration. Certainly raising the levels of public awareness about the importance of hygiene is central to control.

In the UK, government is attempting to increase both the public and the medical profession's concerns by the publication of advice20 and a national plan reviewing the current state of play on healthcare-associated infections and setting out action necessary for their control. Central to this is, of course, the need to reduce the reservoirs of infection, maintaining high standards of hygiene, the prudent use of antimicrobials and the necessary management arrangements to ensure that this occurs. There will be no rapid responses, but this is a long road which must be followed.

## New antibacterials

The 60 years of the ‘antibiotic era’ has been a story of two halves. The first 30 or so years saw the introduction of a multitude of agents, yet in the last 25–30 years, only one new family, the oxazolidinones, has been introduced.

The past few years have seen the major pharmaceutical companies of Aventis, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Roche and Wyeth greatly diminish their research into new antibacterials. It would appear that the antibacterial market has lost its appeal to ‘BigPharma’. The reasons behind this are numerous.21 The market size has remained flat, generics are increasing, pressures against unnecessary use are curtailing prescribing, regulating authorities are increasing the hurdles that new agents have to cross (one wonders if vancomycin or even the aminoglycosides would gain a licence today).

It is calculated that a new agent will cost $300–500 million to develop and ‘BigPharma’ requires such an agent to generate sales of$500–800 million.22 The great hopes of genomics and combinational chemistry have spectacularly failed to produce new antibacterials. So, what of the future? It is to be hoped that the smaller biotechnology companies, with their lesser needs for such high returns in development may produce results. The hope must therefore be for a diverse and thriving smaller bio-technology environment; some of the signs are, however, not encouraging. What has happened with the larger pharmaceutical companies (merging and moving their research base to one country—the USA) is now occurring with the smaller players. For example, Powderject was acquired by Chiron, Amersham by General Electric, and Acambis is moving its research centre to the USA. The UK has only a small percentage of the global biotech revenues (7% as against 73% in the USA) and the capital the UK raised was 1/20 that of the USA.23 One of the few players in the anti-infective area is the UK is Arrow Therapeutics. Government policy should be changed to support and encourage new companies to meet the challenges of this all-important field of therapeutics.

However, the answer might still come from the major companies if and when political pressures increase from the public health needs in response to increasing resistance. In other words, far more, possibly high profile, failures of therapy due to increasing antibacterial resistance will be needed to generate pressure and incentives from government before action is seen, which is not a pleasing prospect.

## Progress

As has been discussed, the main option available to decrease resistance is a decrease in antibiotic usage. Whether this is described as ‘more judicial use’ or an increase in ‘prudent prescribing’, the aim is the same—to decrease the selection pressure on bacteria.

In confined areas of use, such as an ITU, it is not difficult to have a major impact upon either the species of bacteria prevalent at a given time and/or the resistance patterns24,25 but as the clinical conditions require treatment, all that occurs in reality is a ‘squeezing of the balloon’,26 that is, a new bacterium or one with another resistance pattern soon takes over and a new set of problems emerge. However, it does illustrate that a change in prescribing will have a dramatic impact. As we have seen, the majority of prescribing is in the community and one would expect that any change in antibiotic usage might have a more attenuated impact.

There are encouraging signs. Two countries which share certain similarities have pointed the way. In Iceland, where S. pneumoniae resistance to penicillin was genetically linked to resistance to trimethoprim/sulfamethoxazole and the macrolides, a modest 10% reduction in resistance to penicillin was followed by an approximately 30% decrease in use of co-trimoxazole and the macrolides.27 In Finland, erythromycin resistance among Group A streptococci declined from 16.5% in 1992 to 8.6% in 1996, when macrolide consumption declined from 2.4 defined daily doses per 1000 inhabitants in 1991 to 1.38 in 199628 and remained low during the following 5 years. These studies would therefore suggest that a beneficial impact upon the community can be achieved.

The major unknown is the magnitude of such an impact. Mathematical modelling can give some insights and generally shows that resistance can develop rapidly yet decline only slowly after a reduction in use.29,30 The problem can be summed up in Figure 1. The extent and rate of the decline of antibiotic resistance following a major decrease in use will vary and will depend upon the extent of the decrease in consumption, the epidemiological setting (hospital, community, etc.), the organism and, of considerable importance (as seen in Iceland), genetically linked resistance to other agents (and the control of their use).

Antibiotic usage in Europe varies remarkably.31 For example, penicillin use in France is more than four-fold greater than that in the Netherlands, and total antibiotic use in France is more than three times that in the Netherlands and twice that in the UK. I take heart from this knowledge as it suggests that antibiotic use can be reduced in many countries, as there is no suggestion that the population of the Netherlands or the UK have an increased mortality from infections compared with the French.

In the UK, there is some encouraging information: there has been a 21% fall in the numbers of antibiotic prescriptions between 1995 and 2000 in general practice32 (Figure 2). This is good news in itself, but the survey also showed a six-fold variation in the number of prescriptions between different general practices. Again, this suggests there is a major possibility to reduce prescribing even further. However, it is necessary to view all such data with caution, as it is possible that there may be other reasons for a decline in prescribing other than General Practitioner input, such as a decline in, for example, the prevalence of certain common diseases, respiratory tract infections (RTI) for example.

The majority of prescribing is indeed for RTI and these are most commonly encountered in childhood. In England there has been a remarkable reduction in paediatric antibiotic prescribing (Figure 3). There has been a 50% reduction in the use of broad-spectrum penicillins, oral cephalosporins and macrolides since 1993.33 Again this is most encouraging information.

Have these reductions been accompanied by a reduction in the resistance of any important pathogens? As discussed previously, surveillance is very poorly developed, especially in the community, and lacks the basic denominator information. The linkage of usage and resistance rates has, therefore, to be viewed with caution. There is some encouragement however. In England, there has been some suggestion that penicillin resistance (both intermediate and high) in pneumococci has declined, changing from 7.1% in 1999 to 2.8% in 2003.34 Of interest, this has not been accompanied by any significant change in macrolide resistance (14.8% to 12.9%).

## Concluding remarks

Antimicrobial resistance is an inevitable accompaniment to the use of these agents. Profligate use will hasten the decline in their benefit to individuals and society from their lifesaving potential. There is a major need to strengthen our understanding of the dynamics of use and resistance which must be underpinned by robust and meaningful surveillance. Inevitably, society will have to alter its approach to using antimicrobials, and trivial infections must no longer be viewed as a reason for their use. The future is not entirely gloomy since a concerted approach will undoubtedly contribute to preserving these most important therapeutic agents.

Figure 1.

Antibiotic use and resistance—the big unknown.

Figure 1.

Antibiotic use and resistance—the big unknown.

Figure 2.

Number of antibiotic prescriptions in England (in millions) from 1991 to 2000. Data from Ref. 32.

Figure 2.

Number of antibiotic prescriptions in England (in millions) from 1991 to 2000. Data from Ref. 32.

Figure 3.

Number of paediatric formulations of antibiotics prescribed in England from 1993 to 2002. Data from Ref. 33.

Figure 3.

Number of paediatric formulations of antibiotics prescribed in England from 1993 to 2002. Data from Ref. 33.

## References

1.
Denning, D. W., Baily, G. G. & Hood, S. V. (
1997
). Azole resistance in Candida.
European Journal of Clinical Microbiology and Infectious Diseases

16
,
261
–80.
2.
Larder, B. A., Darby, G. & Richman, D. D. (
1989
). HIV with reduced sensitivity to zidovudine (AZT) isolated during prolonged therapy.
Science

243
,
1731
–4.
3.
Wise, R. (
2003
). Antimicrobial resistance: the microbes struggle for survival.
Equine Veterinary Journal

35
,
530
–1.
4.
Darwin, C. (
1859
). On the Origin of Species by Means of Natural Selection. Murray, London.
5.
Jones, J. S. (
1999
). Almost Like a Whale—The Origin of Species Updated, pp.
95
–9. Doubleday, London.
6.
Blazquez, J. (
2003
). Hypermutation as a factor contributing to the acquisition of antibacterial resistance.
Clinical Infectious Diseases

37
,
1201
–9.
7.
Friedland, I. R. & McCracken, G. H. (
1994
). Management of infections caused by antibiotic resistant Streptococcus pneumoniae.
New England Journal of Medicine

331
,
337
–82.
8.
Ho, P. L., Yung, R. W. H., Tsang, D. N. C. et al. (
2001
). Increasing resistance of Streptococcus pneumoniae to fluoroquinolones: results of a Hong Kong multicentre study in 2000.
Journal of Antimicrobial Chemotherapy

48
,
659
–65.
9.
Fraise, A. P., Mitchell, K., O'Brien, S. J. et al. (
1997
). Methicillin-resistant Staphylococcus aureus (MRSA) in nursing homes in a major UK city: an anonymised point prevalence study.
Epidemiology and Infection

118
,
1
–5.
10.
Vandenesch, F., Naimi, T., Enright, M. C. et al. (
2003
). Community-acquired methicillin-resistant Staphylococcus aureus carrying Panton-Valentine Leukocidin genes: worldwide emergence.
Emerging Infectious Diseases

9
,
978
–84.
11.
Baird, D. (
2002
). Extended spectrum β-lactamase producing coliforms—a new problem in a new hospital.
SCIEH Weekly Report

36
,
130
–1.
12.
Fenton, K. A., Ison, C., Johnson, A. P. et al. (
2003
). Ciprofloxacin resistance in Neisseria gonorrhoea in England and Wales in 2002.
Lancet

361
,
1867
–9.
13.
House of Lords. (
1998
). Seventh Report of the House of Lords Select Committee on Science and Technology. Stationery Office, London.
14.
MacGowan, A. P., Bowker, K. E., Bennett, P. M. et al. (
1998
). Surveillance of antimicrobial resistance.
Lancet

352
,
1783
.
15.
Working Party Report. (
2001
). Antimicrobial susceptibility testing.
Journal of Antimicrobial Chemotherapy

48
, Suppl. 1.
16.
Pillay, D. & Zambon, M. (
1998
). Antiviral drug resistance.
British Medical Journal

317
,
660
–2.
17.
Harbarth, S., Harris, A. D., Carmeli, Y. et al. (
2001
). Parallel analysis of individual and aggregated data on antibiotic exposure and resistance in Gram negative bacilli.
Clinical Infectious Diseases

33
,
1462
–8.
18.
House of Lords. (
2003
). Fourth Report of the House of Lords Select Committee on Science and Technology. Stationery Office, London.
19.
Kaplan, S. L., Mason, E. O., Wald, E. R. et al. (
2004
). Decrease in invasive pneumococcal infections in children amongst childrens hospitals in the United States after the introduction of the 7-valent, pneumococcal conjugate vaccine.
Paediatrics

113
,
443
–9.
20.
Department of Health. (
2003
). Winning Ways. Department of Health, London.
21.
Projan, S. J. (
2003
). Why is BigPharma getting out of antibacterial drug discovery?
Current Opinion in Microbiology

6
,
427
–30.
22.
White, R. J. (
2003
). Are the rules different for Biotech companies? In Program and Abstracts of the Forty-third Interscience Conference on Antimicrobial Agents and Chemotherapy, 2003. Abstract, p. 519. American Society for Microbiology, Washington, DC, USA.
23.
Middleton, B. (
2004
). Uphill battle for UK Biotechs.
New Scientist

181
,
52
–3.
24.
Kollef, M. H. (
2001
). Optimizing antibiotic therapy in the intensive care unit setting.
Critical Care

5
,
189
–94.
25.
Loeffler, J. M., Garbino, J., Lew, D. et al. (
2003
). Antibiotic consumption, bacterial resistance and their correlation in a Swiss university hospital and its adult intensive care units.
Scandinavian Journal of Infectious Diseases

35
,
843
–50.
26.
Burke, J. P. (
1999
). Antibiotic resistance—squeezing the balloon?
Journal of the American Medical Association

280
,
1270
–1.
27.
Kristinsson, K. G. (
1997
). Effect of antimicrobial use and other risk factors on antimicrobial resistance in pneumococci.
Microbial Drug Resistance

3
,
117
–23.
28.
Seppala, H., Klaukka, T., Vuopio-Varkila, J. et al. (
1997
). The effect of changes in the consumption of macrolide antibiotics on erythromycin resistance in Group A streptococci in Finland.
New England Journal of Medicine

337
,
441
–6.
29.
Austin, D. J. & Anderson, R. M. (
1999
). Studies of antibiotic resistance within patients, hospitals and the community using simple mathematical models.
Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences

354
,
721
–38.
30.
Austin, D. J., Kristinsson, K. G. & Anderson, R. M. (
1999
). The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance.
Proceedings of the National Academy of Sciences, USA

96
,
800
–1.
31.
ESAC Management Team. Consumption of antibiotics in ambulatory care in Europe. [Online.] http://www.ua.ac.be/esac.
32.
PACT Centre. Antibacterial drugs.http://www.ppa.org.uk/news/pact-092002.htm.
33.
Kendall, H. E., Holdsworth, S. & Sharland, M. (
2004
). GP antibiotic prescribing for children in England from 1993-2002 has fallen by 50%. In 8th Spring Meeting Royal College of Paediatrics and Child Health, York, UK, 2004.
34.
European Antimicrobial Resistance Surveillance System. http://www.earss.rivm.nl.