The Problem of Resistance

What is resistance?

Bacterial isolates have been categorized as being susceptible or resistant to antibiotics ever since they became available. Where the concepts of the minimum inhibitory concentrations (MIC) and minimum bactericidal concentrations of an antibiotic are described. Unfortunately, making an accurate judgement about microbial susceptibility or resistance is somewhat less straightforward than this traditional working definition, since there is usually no simple relationship between the MIC (or minimum bactericidal concentrations) of an antibiotic and clinical response. Therapeutic success depends not only on the concentration of the antibiotic achieved at the site of infection (i.e. its pharmacokinetic behaviour) and its activity against the infecting organisms encountered there (i.e. its pharmacodynamic behaviour), but also on the contribution that the host's own defences are able to make towards clearance of the offending microbes.

The decision as to whether a given bacterial isolate should be labelled susceptible or resistant depends ultimately on the likelihood that an infection with that organism can be expected to respond to treatment with a given drug, but microbiologists and clinicians have become accustomed to the idea that an organism is ‘resistant’ when it is inhibited in vitro by an antibiotic concentration that is greater than that achievable in vivo. Importantly the concentration of antibiotic that is achievable will vary according to the site of infection, dosage, and route of administration. For example, some antibiotics, such as trimethoprim, are excreted primarily via the kidneys and therefore achieve, in the context of urinary tract infections, advantageously high concentrations in urine. Furthermore, the intrinsic activity of an antibiotic against some bacteria (e.g. staphylococci) may be greater than for others (e.g. Escherichia coli) because of the effect of cell envelope structure on achievable intracellular antibiotic concentrations. These issues mean that several different thresholds (breakpoint concentrations) are often used to define susceptibility to an antibiotic. For example, an E. coli strain for which the MIC of ampicillin is 32 mg/l might be classed as susceptible if isolated from a urinary infection, while the same bacterium causing a bloodstream infection would be classified as ampicillin resistant. These differences in definition of susceptibility relate to the variations in achievable concentrations at the site of infection: while an ampicillin concentration of 32 mg/l can reliably be achieved in urine, this is not the case in blood.

Intrinsic resistance

If whole bacterial species are considered, rather than individual isolates, it is apparent immediately that they are not all intrinsically susceptible to all antibiotics (See Table Below); for example, a coliform infection would not be treated with erythromycin, or a streptococcal infection with an aminoglycoside, since the organisms are intrinsically resistant to these antibiotics. Similarly, Pseudomonas aeruginosa and Mycobacterium tuberculosis are intrinsically resistant to most of the agents used to treat more tractable infections. Such intrinsically resistant organisms are sometimes termed non-susceptible, with the term resistant reserved for variants of normally susceptible species that acquire mechanism(s) of resistance.

A microbe will be intrinsically resistant to an antibiotic if it either does not possess a target for the drug's action, or it is impermeable to the drug. Thus, bacteria are intrinsically resistant to polyene antibiotics, such as amphotericin B, as sterols that are present in the fungal but not bacterial cell membrane, are the target for these drugs. The lipopolysaccharide outer envelope of Gram-negative bacteria is important in determining susceptibility patterns, since many antibiotics cannot penetrate this barrier to reach their intracellular target. Fortunately, intrinsic resistance is therefore often predictable, and should not pose problems provided that informed and judicious choices of antibiotics are made for the treatment of infection. Of greater concern is the primarily unpredictable acquisition or emergence of resistance in previously susceptible microbes, sometimes during the course of therapy itself.

Acquired resistance

Introduction of clinically effective antibiotics has been followed invariably by the emergence of resistant strains of bacteria among species that would normally be considered to be susceptible. Acquisition of resistance has seriously reduced the therapeutic value of many important antibiotics, but is also a major stimulus to the constant search for new and more effective antimicrobial drugs. However, while the emergence of resistance to new antibiotics is inevitable, the rate of development and spread of resistance is not predictable.

S, usually considered susceptible; R, usually considered resistant; (S), strain variation in susceptibility; V, variation among related drugs and/or strains.

The first systematic observations of acquired drug resistance were made by Paul Ehrlich between 1902 and 1909 while using dyes and organic arsenicals to treat mice infected experimentally with trypanosomes. Within a very few years of the introduction of sulphonamides and penicillin (in 1935 and 1941 respectively), micro-organisms originally susceptible to these drugs were found to have acquired resistance. When penicillin came into use less than 1% of all strains of Staphylococcus aureus were resistant to its action. By 1946, however, under the selective pressure of this antibiotic, the proportion of penicillin-resistant strains found in hospitals had risen to 14%. A year later, 38% were resistant, and today, resistance is found in more than 90% of all strains of Staph. aureus. In contrast, over the same period, an equally important pathogen, Streptococcus pyogenes, has remained uniformly susceptible to penicillin, although there is no guarantee that resistance will not spread to S. pyogenes in future years.

There is no clear explanation for the marked differences in rate or extent of acquisition of resistance between different species. Possession of the genetic capacity for resistance does not always explain its prevalence in a particular species. Even when selection pressures are similar, the end result may not be the same. Thus, although about 90% of all strains of Staph. aureus are now resistant to penicillin, the same has not happened to ampicillin resistance in E. coli under similar selection pressure. At present, apart from localized outbreaks involving epidemic strains, about 40-50% of E. coli strains are resistant to ampicillin, and this level has remained more or less steady for a number of years. However, since an increasing incidence of resistance is at least partly a consequence of selective pressure, it is not surprising that the withdrawal of an antibiotic from clinical use may often result in a slow reduction in the number of resistant strains encountered in a particular environment. For example, fluoroquinolone resistant strains of P. aeruginosa that emerged in some hospitals as ciprofloxacin or levofloxacin were used more frequently were replaced by more susceptible strains following restriction of removal of these drugs. Conversely, sulphonamide resistant E. coli strains that became commonplace when the sulphonamide-containing combination drug co-trimoxazole was widely used are still prevalent. This is probably because the selection pressure still exists for other antibiotics, such as ampicillin, and the genes coding for sulphonamide and ampicillin resistance are often closely linked on plasmids; hence, use of one antibiotic can select or maintain resistance to another. The introduction of new antibiotics has also resulted in changes to the predominant spectrum of organisms responsible for infections. In the 1960s semi-synthetic ‘β-lactamase stable’ penicillins and cephalosporins were introduced which, temporarily, solved the problem of staphylococcal infections.

Unfortunately, Gram-negative bacteria then became the major pathogens found in hospitals and rapidly acquired resistance to multiple antibiotics in the succeeding years. In the 1970s the pendulum swung the other way with the first outbreaks of hospital infection with multiresistant staphylococci that were resistant to nearly all antistaphylococcal agents. Outbreaks of infection caused by such organisms have occurred subsequently all over the world. There are now signs that Gram-negative bacteria are once again assuming greater importance, particularly in hospitals. Resistance to newer cephalosporins—mediated by extended-spectrum β-lactamases—and fluoroquinolones in E. coli and other enterobacteria is increasing, rendering these commonly used antibiotics less effective. Multiresistant Gram-negative bacteria (such as Acinetobacter species) have emerged that are resistant to most and, occasionally, all approved antibiotics.

Types of acquired resistance

Two main types of acquired resistance may be encountered in bacterial species that would normally be considered susceptible to a particular antibacterial agent.

Mutational resistance

In any large population of bacterial cells a very few individual cells may spontaneously become resistant. Such resistant cells have no particular survival advantage in the absence of antibiotic, but after the introduction of antibiotic treatment susceptible bacterial cells will be killed, so that the (initially) very few resistant cells can proliferate until they eventually form a wholly resistant population. Many antimicrobial agents select for this type of acquired resistance in many different bacterial species, both in vitro and in vivo. The problem has been recognized as being of particular importance in the long-term treatment of tuberculosis with antituberculosis drugs.

Transmissible resistance

A more spectacular type of acquired resistance occurs when genes conferring antibiotic resistance transfer from a resistant bacterial cell to a sensitive one. The simultaneous transfer of resistance to several unrelated antimicrobial agents can be demonstrated readily, both in the laboratory and the patient. Exponential transfer and spread of existing resistance genes through a previously susceptible bacterial population is a much more efficient mechanism of acquiring resistance than the development of resistance by mutation of individual susceptible cells.

Here it is sufficient to stress that however resistance appears in a hitherto susceptible bacterial cell or population, resistance will only become widespread under the selective pressures produced by the presence of appropriate antibiotics. Also, the development of resistant cells does not have to happen often or on a large scale. A single mutation or transfer event can, if the appropriate selective pressures are operating, lead to the replacement of a susceptible population by a resistant one. Without selective pressure, antibiotic resistance may be a handicap rather than an asset to a bacterium.

Cross-resistance and multiple resistance

These terms are often confused. Cross-resistance involves resistance to a number of different members of a group of (usually) chemically related agents that are affected alike by the same resistance mechanism. For example, there is almost complete cross-resistance between the different tetracyclines, because tetracycline resistance results largely from an efflux mechanism that affects all members of the group. The situation is more complex among other antibiotic families. Thus, resistance to aminoglycosides may be mediated by any one of a number of different drug-inactivating enzymes with different substrate specificities, and the range of aminoglycosides to which the organism is resistant will depend on which enzyme it produces. Cross-resistance can also be observed occasionally between unrelated antibiotics. For example, a change in the outer membrane structure of Gram-negative bacilli may concomitantly deny access of unrelated compounds to their target sites. In contrast, multiple drug (multidrug) resistance involves a bacterium becoming resistant to several unrelated antibiotics by different resistance mechanisms. For example, if a staphylococcus is resistant to penicillin, gentamicin, and tetracycline, the resistances must have originated independently, since the strain destroys the penicillin with a β-lactamase, inactivates gentamicin with an aminoglycoside-modifying enzyme, and excludes tetracycline from the cell by an active efflux mechanism.

It is, however, not always clear whether cross-resistance or multiple resistance is being observed. Genes conferring resistance to several unrelated agents can be transferred en bloc from one bacterial cell to another on plasmids, thereby giving the appearance of cross-resistance.

In such cases, detailed biochemical and genetic analysis may be required to prove that the resistance mechanisms are distinct (multiple resistance), although the genes conferring resistance are linked and transferred together on one plasmid.

The clinical problem of drug resistance

Concerns about resistance have been raised at regular intervals since the first introduction of antimicrobial chemotherapy, but awareness of the antibiotic resistance problem has probably never been greater than it is today. It has been suggested that antibiotic resistance is becoming so commonplace that there is a danger of returning to the pre-antibiotic era. It is important not to understate or overstate the problem; the situation is presently becoming serious, but is not yet desperate since most infections are still treatable with several currently available agents. This may, however, mean that the only antibiotics that are still active are more toxic or less effective (or both) than those to which bacteria have acquired resistance. For example, it is generally accepted that glycopeptide antibiotics are less effective in the treatment of Staph. aureus infection than are antistaphylococcal penicillins (e.g. flucloxacillin); since the latter cannot be used against methicillin-resistant Staph. aureus (MRSA), this may partly explain the poorer outcome that is seen in such cases in comparison with infection caused by methicillin-susceptible strains.

There is good evidence that if the antibiotic regimen chosen is subsequently shown to be inactive against the pathogens causing infection, then patient outcome is worse (See Fig. Below). This means that clinicians are likely to opt for unnecessarily broad-spectrum therapy particularly in critically ill patients. Unfortunately, repeated use of such regimens against bacteria that harbour resistance genes intensifies the selective pressure for further resistance development, notably in hospital, where the most vulnerable patients are managed.

In many less-developed countries of the world the therapeutic options may be severely restricted for economic reasons. There is no doubt that the problem of antibiotic resistance is a global issue, and in future years there is a real possibility that physicians will be faced increasingly with infections for which effective treatment is not available. Some of the organisms in which resistance is a particular problem are summarized below.

Enteric Gram-negative bacteria

The prevalence of resistance in hospital strains of enteric Gram-negative bacteria has been rising steadily for the past 40 years, particularly in large units. Although cephalosporins, quinolones, and aminoglycosides have been developed to cope with the problem, resistance to these newer compounds continues to increase in most countries. Outbreaks of infection caused by multiresistant Klebsiella strains and extended-spectrum β-lactamase-producing enterobacteria in general are being reported with worrying frequency, especially in high dependency areas of hospitals.

Mortality recorded in three separate studies for patients who received antibiotic treatment that was subsequently shown to be inactive (dark grey) or active (light grey) against pathogens isolated. VAP, ventilator associated pneumonia. Data from: Kollef MH, Ward S. The influence of mini-BAL cultures on patient outcomes: implications for the antibiotic management of ventilator-associated pneumonia. Chest 1998; 113: 412-420; Kollef MH Sherman G, Ward S, Fraser VJ Inadequate antimicrobial treatment of infections: a risk factor for hospital mortality among critically ill patients. Chest 1999; 115: 462-474; Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH. The influence of inadequate antimicrobial treatment of bloodstream infections on patient outcomes in the ICU setting. Chest 2000; 118: 146-155.

Widespread resistance in enteric bacteria is a particular problem in less-developed areas of the world where heavy and indiscriminate use of antibiotics may combine with a high prevalence of drug-resistant bacteria in the faecal flora, poor standards of sanitation, and a high incidence of diarrhoeal disease to encourage the rapid emergence and spread of multiresistant strains of enteric bacteria. Epidemics of diarrhoeal disease caused by multiresistant strains of intestinal pathogens, including Vibrio cholerae, shigellae, salmonellae, and toxin-producing strains of E. coli, have occurred around the world.


Acinetobacter

These organisms cause hospital-acquired infections especially in patients in intensive care units, e.g. ventilator-associated pneumonia. Such infections are usually extremely difficult to treat because of the multiple classes of antibiotic resistance found in these bacteria. Very few antibiotics are now reliably effective for treatment of acinetobacter infections. Even with carbapenems such as imipenem, resistance has started to emerge. Colistin, a relatively toxic old antibiotic that has largely been abandoned for systemic administration, has been used to treat some strains that were resistant to all other licensed antibiotics. Many multiresistant Acinetobacter spp. strains are currently susceptible to the new antibiotic tigecycline, and this may be helpful in such infections.

Staphylococci and enterococci

MRSA is endemic in many hospitals and nursing homes. The proportion of Staph. aureus isolates causing serious sepsis, such as bloodstream infection, that are resistant to methicillin has reached 40-50% in the UK and some countries in southern Europe. The prevalence is even higher in countries in the Far East and USA. MRSA infections have often been treated with glycopeptides, and isolates with low-level resistance to these antibiotics can be found. Very occasional MRSA strains with high-level resistance to glycopeptides have also been reported. Several newer antibiotics, including linezolid, daptomycin, and tigecycline, are active against these strains, but occasional reports of resistance have already occurred.

Coagulase-negative staphylococci and enterococci are often multiresistant and cause infections typically in patients with indwelling prosthetic material, such as catheters, vascular grafts, joints and heart valves. A combination of antibiotics may be required to treat serious enterococcal infections, but the emergence of high-level aminoglycoside resistance may seriously limit this option. Enterococci carrying genes conferring high-level resistance to glycopeptides have emerged.

Linezolid has been used successfully to treat infection caused by such strains, but resistance has some occurred in patients receiving long courses of therapy, particularly if the focus of infection has not been removed.

Streptococcus pneumoniae

Another major problem concerns the emergence of resistance in Streptococcus pneumoniae, the most common cause of community-acquired pneumonia and other respiratory infections. This organism used to be combated easily by treatment with penicillin and its derivatives. Unfortunately, isolates with resistance to most antibiotics can now be found in most countries of the world. Such infections are often treated with broad-spectrum cephalosporins, which can attain sufficient tissue concentrations to exceed the raised MIC for these strains. The prevalence of macrolide-resistant pneumococci tends to correlate with how often these antibiotics are used, especially in the community where most respiratory tract infections are treated. Newer fluoroquinolones such as moxifloxacin have increased activity against pneumococci. Some resistance emergence has developed in units where these agents have been used commonly.

Neisseria meningitidis

Decreased levels of susceptibility to penicillin have been seen in many countries, but high-level resistance is exceptionally rare. The emergence of resistance to penicillin in N. meningitidis has important strategic implications because of the need for immediate treatment of the life-threatening infections caused by these organisms. Currently penicillin is still used empirically in some cases of suspected meningococcal infection. The cephalosporins cefotaxime or ceftriaxone are often favoured for the empirical treatment of meningitis because the antibiotic concentration achieved in the cerebrospinal fluid more reliably exceeds the MIC for the pathogen in both meningococcal and pneumococcal infection.

Tuberculosis

Strains of M. tuberculosis that are resistant to two or more of the first line drugs—isoniazid, ethambutol, rifampicin, and streptomycin—are increasing common, particularly in HIV-infected patients. The prevalence of resistant strains varies markedly between countries: in 0-14% of new cases of tuberculosis (median: 1.4%) and 0-54% of previously treated cases (median: 13%) in recent World Health Organization surveys. The resistant bacteria can be transmitted, for example in hospitals, prisons and in the community, and represent a major public health issue. The emergence of resistance is associated with poor compliance with antituberculosis medication. Directly observed therapy is increasingly advocated therefore for patients in whom compliance may be unreliable.