Chapter 23: Antibacterial Agents and Resistance

The mode of action of antimicrobials on bacteria is the focus of this chapter. Resistance to antibacterial agents, and strategies to minimize resistance, are also addressed here. Specific information about pathogenic bacteria can be found in Chapters 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41; a complete guide to the treatment of infectious diseases is beyond the scope of this book.

Natural materials with some activity against microbes were used in folk medicine in earlier times, such as the bark of the cinchona tree (containing quinine) in the treatment of malaria. Rational approaches to chemotherapy began with Ehrlich’s development of arsenical compounds for the treatment of syphilis early in the 20th century. Many years then elapsed before the next major development, which was the discovery of the therapeutic effectiveness of a sulfonamide (prontosil rubrum) by Domagk in 1935. Penicillin had been discovered in 1929 by Fleming, but could not be adequately purified at that time; this was accomplished later, and penicillin was produced in sufficient quantities so that Florey and colleagues could demonstrate its clinical effectiveness in the early 1940s.

Since that time, numerous new antimicrobial agents have been discovered or developed, and many have found their way into clinical practice. Thanks to these medicines, the human experience in industrialized nations is dramatically different today than it was in the preantibiotic era. However, this success has come at the cost of rising antimicrobial resistance. In order to be good antimicrobial stewards, all clinicians must understand the ways in which these drugs work, the ways in which bacteria evolve in response to antibiotics, and strategies for their judicious use.

GENERAL CONSIDERATIONS

Clinically effective antimicrobial agents exhibit selective toxicity toward the microbe rather than the host, a characteristic that differentiates them from the disinfectants (see Chapter 3). In most cases, selectivity is explained by action on microbial processes or structures that differ from those of mammalian cells. For example, some agents inhibit the synthesis of the bacterial cell wall (an organelle not present in eukaryotes), and others act on the 70S bacterial ribosome (but not the 80S eukaryotic ribosome). Some antimicrobials, such as penicillin, are usually nontoxic to the host, unless hypersensitivity develops. For others, such as the aminoglycosides, the effective therapeutic dose is relatively close to the toxic dose; as a result, control of dosage and blood levels must be much more precise.

Definitions

• Antibiotics—antimicrobials of microbial origin, most of which are produced by fungi or by bacteria of the genus Streptomyces.

• Antimicrobials—substances used in the treatment of infectious diseases, including antibiotics and other antibacterials, antifungals, antiparasitics, and antivirals.

• Bactericidal—antimicrobial activity that not only inhibits growth but is lethal to bacteria.

• Bacteriostatic—antimicrobial activity that inhibits growth but does not kill the organisms. The host defense mechanisms are ultimately responsible for eradication of infection.

• Minimal inhibitory concentration (MIC)—a laboratory term that defines the lowest concentration (μg/mL) able to inhibit growth of the microorganism in vitro.

• Resistant, nonsusceptible—terms applied when organisms are not inhibited by clinically achievable concentrations of an antimicrobial agent.

• Sensitive, susceptible—terms applied to microorganisms indicating that they will be inhibited by concentrations of the antimicrobial that can be achieved clinically.

• Spectrum—an expression of the categories of microorganisms against which an antimicrobial is typically active. A narrow-spectrum agent has activity against only a few organisms. A broad-spectrum agent has activity against diverse types of organisms (eg, both gram-positive and gram-negative bacteria).

Sources of Antimicrobial Agents

There are three main sources of antimicrobial agents.

First are antibiotics, which are molecules of biological origin. They probably play an important part in microbial ecology in the natural environment. Penicillin, for example, is produced by several molds of the genus Penicillium, and the first cephalosporin antibiotics were derived from other molds. Another source of naturally occurring antibiotics is the genus Streptomyces, which are gram-positive, branching bacteria found in soil and freshwater sediments. Streptomycin, the tetracyclines, chloramphenicol, erythromycin, and many other antibiotics were discovered by screening large numbers of Streptomyces isolates from different parts of the world. Antibiotics are mass produced by techniques derived from the procedures of the fermentation industry.

Second are the chemically synthesized antimicrobial agents. These were initially discovered among compounds synthesized for other purposes and tested for their therapeutic effectiveness in animals. The sulfonamides, for example, were discovered as a result of routine screening of aniline dyes. More recently, active compounds have been synthesized with structures tailored to be effective inhibitors or competitors of known metabolic pathways. Trimethoprim, which inhibits dihydrofolate reductase, is an excellent example. “Structure-based drug design” involves the use of X-ray crystallography to understand the three-dimensional molecular conformation of potential drug targets, and then synthesizing small molecules to bind those targets. This technique holds great promise, although relatively few antimicrobials have yet been developed in this manner.

A third source of antimicrobials arises from the molecular manipulation of previously discovered antibiotics to broaden their range and degree of activity against microorganisms or to improve their pharmacologic characteristics. Examples include the development of penicillinase-resistant and broad-spectrum penicillins, as well as a large range of aminoglycosides and cephalosporins of increasing activity, spectrum, and resistance to inactivating enzymes.

Spectrum of Action

The spectrum of activity of each antimicrobial agent describes the genera and species against which it is typically active. See Table 23–1 for the most common antimicrobial agents and bacteria. Spectra overlap but are usually characteristic for each broad class of antimicrobial. Some antibacterial antimicrobials are known as narrow-spectrum agents; for example, benzyl penicillin is highly active against many gram-positive and gram-negative cocci but has little activity against enteric gram-negative bacilli. The tetracyclines, the cephalosporins, and the carbapenems, on the other hand, are broad-spectrum agents that inhibit a wide range of gram-positive and gram-negative bacteria.

SELECTED ANTIBACTERIAL AGENTS

The major antimicrobials are now considered in more detail, with emphasis on their modes of action and spectrum. Details on specific antimicrobial agent use, dosage, and toxicity should be sought in a specialized text or handbook written for that purpose.

Antimicrobials That Act on Cell Wall Synthesis

Without an intact wall, these cells become fragile, subject to osmotic stress, and are more easily eliminated by the immune system. The peptidoglycan component of the bacterial cell wall provides its shape and rigidity. This giant molecule is formed by weaving the linear glycans N-acetylglucosamine and N-acetylmuramic acid into a basket-like structure. Mature peptidoglycan is held together by cross-linked short peptide side chains hanging off the long glycan molecules. This cross-linking process is the target of two of the most important groups of antimicrobials, the β-lactams and the glycopeptides (including vancomycin) (Figure 23–1). Peptidoglycan is unique to bacteria and its synthesis is described in more detail in Chapter 21.

BETA-Lactam Antimicrobials

The β-lactam antimicrobial agents comprise the penicillins, cephalosporins, carbapenems, and monobactams. They are named after the β-lactam ring in their structure; this ring is essential for their antibacterial activity. Penicillin, the first member of this class, was derived from molds of the genus Penicillium. Later β-lactams were derived from both molds and bacteria of the genus Streptomyces. Today it is possible to synthesize β-lactams, but most are derived from semisynthetic processes involving chemical modification of the products of fermentation.

The β-lactam antibacterial agents interfere with the transpeptidation reactions that seal the peptide crosslinks between glycan chains. They do so by interference with the action of the transpeptidase enzymes which carry out this cross-linking. These targets of all the β-lactams are commonly called penicillin-binding proteins (PBPs). Several distinct PBPs occur in any one strain and vary in their avidity of binding to different β-lactam drugs.

The β-lactams are classified by chemical structure (Figure 23–2). They may have one β-lactam ring (monobactams), or a β-lactam ring fused to a five-member thiazolidine penem ring (penicillins, carbapenems) or a six-member dihydrothiazine cephem ring (cephalosporins). Within these major groups, differences in the side chain(s) attached to the single or double ring can have a significant effect on the drug’s pharmacologic properties and spectrum. These properties include resistance to gastric acid, which allows oral administration and their pattern of distribution into body compartments (eg, blood, cerebrospinal fluid, joints). The features that alter their spectrum include permeability into the bacterial cell, affinity for PBPs, and vulnerability to the various bacterial mechanisms of resistance.

β-Lactam antimicrobials are usually highly bactericidal, but only to growing bacteria synthesizing new cell walls. Killing involves attenuation and disruption of the developing peptidoglycan “corset,” liberation or activation of autolytic enzymes that further disrupt weakened areas of the wall, and finally osmotic lysis due to passage of water through the cytoplasmic membrane to the hypertonic interior of the cell. As might be anticipated, cell wall-deficient organisms, such as Mycoplasma, are not susceptible to β-lactam antimicrobials.

Penicillins

Penicillin G is the oldest penicillin. It remains active primarily against certain gram-positive organisms, many gram-negative cocci, and some spirochetes, including Treponema pallidum, the cause of syphilis. It has little action against most gram-negative bacilli, because their outer membrane prevents passage of these antibiotics to their sites of action on cell wall synthesis. Penicillin G is the least toxic of the penicillins. Its modification as penicillin V confers acid stability, so it can be given orally.

Three major strategies in drug development have allowed penicillins to remain an important antibiotic class. First, semisynthetic penicillins were developed to cope with staphylococcal penicillinase. This penicillinase is one of a family of bacterial enzymes called β-lactamases that inactivate β-lactam antimicrobials. The penicillinase-resistant penicillins (methicillin, nafcillin, oxacillin, dicloxacillin) have narrow spectra, but are active against penicillinase-producing Staphylococcus aureus (although methicillin is no longer in use, these bacteria are still commonly referred to as “methicillin-susceptible S aureus, or “MSSA”).

Second, a group of broader spectrum penicillins was created, which owe their expanded activity to their ability to traverse the outer membrane of some gram-negative bacteria, and in some cases to their resistance to hydrolysis by gram-negative β-lactamases. Some, such as the aminopenicillins ampicillin and amoxicillin, have excellent activity against a range of gram-negative pathogens but not against Pseudomonas aeruginosa, an important opportunistic pathogen. Others, such as the ureidopenicillin piperacillin, are active against Pseudomonas when given in high dosage. These penicillins with a gram-negative spectrum are slightly less active than penicillin G against gram-positive organisms and are inactivated by staphylococcal penicillinase. Finally, in order to combat bacterial β-lactamases, penicillins are sometimes dosed with β-lactamase inhibitors (see later).

Cephalosporins

The structure of the cephalosporins confers resistance to hydrolysis by staphylococcal penicillinase and to varying degrees the β-lactamases of groups of gram-negative bacilli. The cephalosporins are classified by generation—first, second, third, fourth, fifth, or “unclassified.” The “generation” term relates to historical breakthroughs in expanding their spectrum through modification of the side chains. In general, a cephalosporin of a higher generation has a wider spectrum, and in some instances, more quantitative activity (lower MIC) against gram-negative bacteria. As the gram-negative spectrum increases, these agents typically lose some of their potency (higher MIC) against gram-positive bacteria.

The first-generation cephalosporins cefazolin and cephalexin have a spectrum of activity against gram-positive organisms that resembles that of the penicillinase-resistant penicillins. In addition, they are active against some of the Enterobacteriaceae (see Table 23–1). These agents continue to have therapeutic value because of their high activity against gram-positive organisms, because they are well tolerated, and because a broader spectrum is unnecessary in many infections due to methicillin susceptible staphylococci and streptococci.

Second-generation cephalosporins such as cefoxitin and cefaclor are resistant to β-lactamases of some gram-negative organisms that inactivate first-generation compounds. Of particular importance is their expanded activity against Enterobacteriaceae species, although in theory this comes at the cost of reduced effectiveness against certain gram positives.

Third-generation cephalosporins, such as ceftriaxone, cefotaxime, and ceftazidime, have an even wider spectrum; they are active against gram-negative organisms, often at MICs that are 10- to 100-fold lower than first-generation compounds. Of these three agents, only ceftazidime is active against P aeruginosa. The potency, broad spectrum, and low toxicity of the third-generation cephalosporins have made them preferred agents in life-threatening infections in which the causative organism has not yet been isolated. Selection depends on the clinical circumstances. For example, ceftriaxone or cefotaxime are preferred for meningitis because they have the highest activity against the three major causes, Neisseria meningitidis, Streptococcus pneumoniae, and Haemophilus influenzae. For a febrile stem cell transplant patient, ceftazidime might be chosen because of the higher likelihood of P aeruginosa involvement.

Fourth-generation cephalosporins retain much of the gram-positive coverage of ceftriaxone, and have enhanced ability to cross the outer membrane of gram-negative bacteria as well as resistance to many gram-negative β-lactamases. Compounds such as cefepime have activity against a wider spectrum of Enterobacteriaceae as well as P aeruginosa. These cephalosporins retain the high affinity of third-generation drugs and activity against Neisseria and H influenzae. In effect, these drugs can be conceptualized as having the activity of ceftriaxone plus ceftazidime.

Ceftaroline has the unique ability to bind avidly to PBP-2A, the altered penicillin-binding protein that confers resistance to other β-lactam antibiotics in methicillin-resistant S aureus (MRSA). See Chapter 24 for information on MRSA. It retains some activity against Enterobacteriaceae, although it should be thought of primarily as an anti–gram-positive agent. For this reason, some prefer to categorize ceftaroline as “unclassified” rather than as “fifth generation,” because it bucks the trend of adding gram-negative coverage as is seen when comparing fourth-generation cephalosporins to first generation ones. In effect, think of ceftaroline as having a similar spectrum of activity as ceftriaxone plus coverage of MRSA.

Fifth-generation cephalosporins such as ceftolozane are built to kill highly drug-resistant gram-negative bacteria, including P aeruginosa. Ceftolozane is dosed in combination with the β-lactamase inhibitor tazobactam to partially protect it from the activity of these hydrolytic enzymes (see section on β-lactamase inhibitors). Ceftolozane has less activity against gram-positive bacteria than earlier generation cephalosporins.

Carbapenems

The carbapenems imipenem, meropenem, and doripenem have the broadest spectrum of all β-lactam antibiotics. This fact appears to be due to the combination of easy penetration of gram-negative and gram-positive bacterial cells and high level of resistance to β-lactamases. All three agents are active against streptococci, retain some antistaphylococcal activity, and are highly active against both β-lactamase-positive and negative strains of gonococci and H influenzae. In addition, they are as or more active than third-generation cephalosporins against gram-negative rods. They are highly effective against obligate anaerobes such as Bacteroides fragilis. A closely related drug, ertapenem, is ineffective against Pseudomonas, but is otherwise similar. Imipenem is the carbapenem of choice against gram-positive pathogens, but it is rapidly hydrolyzed by renal tubular dehydropeptidase-1; therefore, it is administered together with an inhibitor of this enzyme (cilastatin), which greatly improves its urine levels and other pharmacokinetic characteristics. Meropenem, doripenem, and ertapenem are not significantly degraded by dehydropeptidase-1 and do not require coadministration of cilastatin. None of the carbapenems available in the United States today are administered orally.

Monobactams

Aztreonam, the first monobactam licensed in the United States, has a spectrum limited to aerobic and facultatively anaerobic gram-negative bacteria, including Enterobacteriaceae, P aeruginosa, Haemophilus, and Neisseria. Monobactams have poor affinity for the PBPs of gram-positive organisms and strict anaerobes and thus demonstrate little activity against them. However, they are highly resistant to hydrolysis by β-lactamases of gram-negative bacilli. Anaerobic superinfections and major distortions of the bowel flora are less common with aztreonam therapy than with other broad-spectrum β-lactam antimicrobials, presumably because aztreonam does not produce a general suppression of gut anaerobes.

β-Lactamase Inhibitors

A number of β-lactams with little or no antimicrobial activity are capable of binding irreversibly to β-lactamase enzymes and, in the process, rendering them inactive. Three such compounds, clavulanic acid, sulbactam, and tazobactam, are referred to as suicide inhibitors, because they must first be hydrolyzed by a β-lactamase before becoming effective inactivators of the enzyme. They are highly effective against staphylococcal penicillinases and broad-spectrum β-lactamases; however, their ability to inhibit cephalosporinases is significantly less. Combinations of one of these inhibitors with an appropriate β-lactam antimicrobial agent protects the therapeutic agent from destruction by many β-lactamases and significantly enhances its spectrum. Four such combinations are now available in the United States: amoxicillin/clavulanate, ampicillin/sulbactam, ceftolozane/tazobactam, and piperacillin/tazobactam. Bacteria that produce chromosomally encoded inducible cephalosporinases are not susceptible to these combinations. A non–β-lactam β-lactamase inhibitor, avibactam, is now available in combination with ceftazidime, and this agent may be combined with carbapenems in the near future. Avibactam inhibits these hydrolytic enzymes indirectly, rather than via suicide inhibition as with the other inhibitors. It provides superior blockade of certain β-lactamases found in difficult-to-treat gram-negative infections, such as those caused by KPC and ampC producers. See the following section on enzymatic inactivation as a resistance mechanism for more information on β-lactam resistance.

Clinical Use

The β-lactam antibiotics are usually the drugs of choice for infections caused by susceptible organisms because of their bactericidal action and low toxicity. They also have great value in the prevention of many infections, such as surgical site infections. Most are excreted by the kidney and achieve high urinary levels. Penicillins reach the cerebrospinal fluid when the meninges are inflamed and are effective in the treatment of meningitis, although first- and second-generation cephalosporins are not. In contrast, the third-generation cephalosporins penetrate the blood-brain barrier much better and have become the agents of choice in the treatment of undiagnosed meningitis and meningitis caused by most gram-negative organisms.

Glycopeptide Antimicrobials

Vancomycin and teicoplanin belong to this group. Each of these antimicrobials inhibits assembly of the linear peptidoglycan molecule by binding directly to the terminal amino acids of the peptide side chains. The effect is the same as with β-lactams: prevention of peptidoglycan cross-linking. Both agents are bactericidal, but are primarily active only against gram-positive bacteria. Their main use has been against multidrug-resistant gram-positive infections including those caused by strains of staphylococci that are resistant to the penicillinase-resistant penicillins and most cephalosporins, especially MRSA. Neither agent is absorbed by mouth; this feature allows these medications to be given orally to treat Clostridium difficile infections of the bowel (see Chapter 29). A related drug, telavancin, was created by adding a lipid tail onto a glycopeptide backbone, thus giving it the theoretical advantage of cell membrane activity and cell wall activity; its clinical usefulness remains to be firmly established. Semi-synthetic glycopeptides such as dalbavancin and oritavancin have been modified to greatly increase their half lives, such that they may be dosed once per week for certain infections.

KEY CONCLUSIONS

Inhibitors of Protein Synthesis

(Figure 23–3)

Aminoglycosides

All members of the aminoglycoside group of antibacterial agents have a six-member aminocyclitol ring with attached amino sugars. The individual agents differ in terms of the exact ring structure and the number and nature of the amino sugar residues. Aminoglycosides are active against a wide range of bacteria, but only those organisms that are able to transport them into the cell by a mechanism that involves oxidative phosphorylation. Thus, they have little or no activity against strict anaerobes or facultative organisms that metabolize only fermentatively (eg, streptococci). It appears highly probable that aminoglycoside activity against facultative organisms is similarly reduced in vivo when the oxidation–reduction potential is low, as in abscesses.

Once inside bacterial cells, aminoglycosides inhibit protein synthesis by binding to the bacterial ribosomes either directly or by involving other proteins. This binding destabilizes the ribosomes and blocks initiation complexes, thus preventing the elongation of polypeptide chains. The agents may also cause distortion of the site of attachment of mRNA, mistranslation of codons, and failure to produce the correct amino acid sequence in proteins. The first aminoglycoside, streptomycin, binds to the 30S ribosomal subunit, but the newer and more active aminoglycosides bind to multiple sites on both 30S and 50S subunits. This gives the newer agents broader spectra and less susceptibility to resistance caused by binding site mutation.

Eukaryotic ribosomes are resistant to aminoglycosides, and the antimicrobials are not actively transported into eukaryotic cells. These properties account for their selective toxicity and also explain their ineffectiveness against intracellular bacteria such as Rickettsia and Chlamydia.

Gentamicin and tobramycin are the major aminoglycosides; they have an extended spectrum, which includes Enterobacteriaceae, and of particular importance, P aeruginosa. They are sometimes beneficial in treating serious infections caused by gram-positive pathogens such as S aureus and enterococci, but only when combined with other drugs. Streptomycin and amikacin are now primarily used in combination with other antimicrobial agents in the therapy of tuberculosis and other mycobacterial diseases. Neomycin, the most toxic aminoglycoside, is used in topical preparations and as an oral preparation before certain types of intestinal surgery, because it is poorly absorbed.

All of the aminoglycosides are toxic to the vestibular and auditory branches of the eighth cranial nerve to varying degrees; this damage can lead to complete and irreversible loss of hearing and balance. These agents may also be toxic to the kidneys. It is often essential to monitor blood levels during therapy to ensure adequate yet nontoxic doses, especially when renal impairment diminishes excretion of the drug. Patients with cystic fibrosis may benefit from inhaled tobramycin when treating P aeruginosa lung infections because high concentrations are achieved and there is little if any absorption into the bloodstream in these patients, thus reducing toxicity risk.

The clinical value of the aminoglycosides is a consequence of their rapid bactericidal effect, their broad spectrum, and the slow development of bacterial resistance, including retained action against Pseudomonas strains that resist many other drugs. They cause fewer disturbances of the resident microbiota than most other broad-spectrum antimicrobials, probably because of their lack of activity against the predominantly anaerobic flora of the bowel, and because they are only used parenterally for systemic infections. The β-lactam antibiotics often act synergistically with the aminoglycosides, most likely because their action on the cell wall facilitates aminoglycoside penetration into the bacterial cell. This effect is most pronounced with organisms such as streptococci and enterococci, which lack the metabolic pathways required to transport aminoglycosides to their interior.

Tetracyclines

Tetracyclines are composed of four fused benzene rings. Substitutions on these rings provide differences in pharmacologic features of the major members of the group, doxycycline and minocycline. The tetracyclines inhibit protein synthesis by binding to the 30S ribosomal subunit at a point that blocks attachment of aminoacyl-tRNA to the acceptor site on the mRNA ribosome complex. Unlike the aminoglycosides, their effect is reversible. They are bacteriostatic rather than bactericidal. Tigecycline belongs to a related class, the glycylcyclines. It covers anaerobes aggressively and thus may be used for treating polymicrobial intraabdominal infections and other complicated deep-tissue infections. Because it is poorly tolerated from a gastrointestinal standpoint, and because of concerns for clinical failure when used for bloodstream infections, this drug’s most useful role may be in the treatment of nontuberculous mycobacterial infections.

The tetracyclines are broad-spectrum agents with a range of activity that encompass most common pathogenic species, including gram-positive and gram-negative rods and cocci and both aerobes and anaerobes. They are active against cell wall-deficient organisms, such as Mycoplasma, and against some obligate intracellular bacteria, including members of the genera Rickettsia and Chlamydia. Differences in spectrum of activity between members of the group are relatively minor. Acquired resistance to one generally confers resistance to all; however, tigecycline appears to overcome the major resistance mechanisms to other tetracyclines, and thus may be a useful alternative in select cases.

The tetracyclines are absorbed orally, whereas tigecycline is not. Tetracyclines are chelated by divalent cations, which may reduce their absorption and activity. Thus, they should not be taken with dairy products or many antacid preparations. Tetracyclines are excreted in the bile and urine in active form.

The original tetracycline drug had a strong affinity for developing bone and teeth, to which it gave a yellowish color and enamel damage, and thus it was avoided in children up to 8 years of age. This is less of a significant problem with doxycycline, however. For life-threatening infections such as Rocky Mountain spotted fever, patients should be treated with doxycycline regardless of their age. Common complications of tetracycline therapy include photosensitivity, nausea, and esophagitis.

Chloramphenicol

Chloramphenicol has a simple nitrobenzene ring structure that can be mass produced by chemical synthesis. It influences protein synthesis by binding to the 50S ribosomal subunit and blocking the action of peptidyl transferase, which prevents formation of the peptide bond essential for extension of the peptide chain. Its action is reversible in most susceptible species; thus, it is bacteriostatic. It has little effect on eukaryotic ribosomes, which explains its selective toxicity.

Like tetracycline, chloramphenicol is a broad-spectrum antibiotic with a wide range of activity against both aerobic and anaerobic species (see Table 23–1). Chloramphenicol is readily absorbed from the upper gastrointestinal tract and diffuses readily into most body compartments, including the cerebrospinal fluid. It also permeates readily into mammalian cells and is active against obligate intracellular pathogens such as Rickettsia and Chlamydia. It is poorly concentrated in urine.

The major drawback to this inexpensive, broad-spectrum antimicrobial with almost ideal pharmacologic features is a rare but serious toxicity. Between 1 in 100 000 and 1 in one million patients treated with even low doses of chloramphenicol have an idiosyncratic reaction that results in aplastic anemia. The condition is irreversible and, before the advent of stem cell transplantation, was universally fatal. In high doses, chloramphenicol also causes a reversible depression of the bone marrow and, in neonates may cause abdominal, circulatory, and respiratory dysfunction. The inability of the immature infant liver to conjugate and excrete chloramphenicol aggravates this latter condition.

In the United States, chloramphenicol use is now restricted to the treatment of rickettsial or ehrlichial infections in which tetracyclines are relatively contraindicated because of hypersensitivity or pregnancy. In some developing countries, chloramphenicol is used more extensively because of its low cost and proven efficacy in diseases such as typhoid fever and bacterial meningitis.

A related class of drugs, the pleuromutilins, also blocks peptidyl transferase at the 50S ribosomal subunit. Retapamulin is used topically for relatively superficial streptococcal and staphylococcal skin infections.

Macrolides

The macrolides erythromycin, azithromycin, and clarithromycin, differ in their composition of a large 14- or 15-member ring structure. They affect protein synthesis at the ribosomal level by binding to the 50S subunit and blocking the translocation reaction. Their effect is primarily bacteriostatic. Macrolides, which are concentrated in phagocytes and other cells, are effective against some intracellular pathogens.

Erythromycin, the first macrolide, has a spectrum of activity that includes many pathogenic gram-positive bacteria and some gram-negative organisms. Its gram-negative spectrum includes Neisseria, Bordetella, Campylobacter, and Legionella, but not the Enterobacteriaceae. Erythromycin and related drugs are also effective against Chlamydia and Mycoplasma.

Bacteria that have developed resistance to erythromycin are usually resistant to the newer macrolides azithromycin and clarithromycin as well. These newer agents have the same spectrum as erythromycin, with some significant additions. Azithromycin has quantitatively greater activity (lower MICs) against most of the same gram-negative bacteria. Clarithromycin is the most active of the three against both gram-positive and gram-negative pathogens, and it is also active against mycobacteria. In addition, both azithromycin and clarithromycin have demonstrated efficacy against Borrelia burgdorferi, the causal agent of Lyme disease and the protozoan parasite Toxoplasma gondii, which causes toxoplasmosis. Both azithromycin and clarithromycin may have undesirable side effects, including GI upset and cardiac arrhythmias. Azithromycin may benefit patients with cavitary lung disease due in part to its anti-inflammatory effects. A related drug, telithromycin, belongs to the ketolide class; it is less susceptible to bacterial resistance mechanisms, but has been associated with liver toxicity and is thus rarely used.

Clindamycin

Clindamycin is a lincosamide, chemically unrelated to the macrolides but with a similar mode of action and spectrum. It has greater activity than the macrolides against gram-negative anaerobes, including the important B fragilis group. Although clindamycin is a perfectly adequate substitute for a macrolide in many situations, its primary use is in instances where anaerobes are or may be involved. In addition, there is experimental evidence that clindamycin may mitigate toxin production by highly virulent S aureus and Streptococcus pyogenes strains. For this reason, many clinicians add it to a bactericidal agent such as nafcillin or vancomycin for treatment of serious deep-tissue infections caused by these organisms.

Oxazolidinones

Linezolid is the most widely used of a class of antibiotics that act by binding to the bacterial 50S ribosome of gram-positive organisms, and many mycobacteria and anaerobes. It does not cover aerobic gram negatives. Oxazolidinones are clinically useful in pneumonia and soft tissue infections, particularly those caused by resistant strains of staphylococci and enterococci. Risk of bone marrow suppression is notorious for linezolid, especially when dosed for more than 2 weeks. Optic and peripheral neuropathy have been reported. It is also a monoamine oxidase inhibitor, and thus may precipitate a systemic reaction called the serotonin syndrome when given to patients simultaneously taking certain antidepressants. Tedizolid, a related drug, may have a lower risk of this complication.

Streptogramins

Quinupristin and dalfopristin are used in a synergistic combination known as synercid. They inhibit protein synthesis by binding to different sites on the 50S bacterial ribosome of certain gram positives, including MRSA and vancomycin-resistant enterococci (VRE); quinupristin inhibits peptide chain elongation, and dalfopristin interferes with peptidyl transferase. Muscle pain is a common and often-limiting side effect. Their clinical use thus far has been limited generally to the treatment of VRE.

KEY CONCLUSIONS

Inhibitors of Nucleic Acid Synthesis

(Figure 23–4)

Quinolones

The quinolones have a nucleus of two fused six-member rings that when substituted with fluorine become fluoroquinolones, which are now the dominant quinolones for the treatment of bacterial infections. The fluoroquinolones now in use are ciprofloxacin, gemifloxacin, moxifloxacin, and levofloxacin. The addition of a piperazine ring and its methylation alter the activity and pharmacologic properties of each individual compound. The target of the quinolones are DNA gyrase and topoisomerase IV, the enzymes responsible for nicking, supercoiling, and sealing bacterial DNA during replication. Binding to two enzymes reduces the chance a single mutation can lead to resistance, which was a problem with the first quinolone, nalidixic acid, a single binding-site agent.

The fluoroquinolones are highly active and bactericidal against a wide range of aerobes and facultative anaerobes. However, strict anaerobes are generally resistant. Levofloxacin and moxifloxacin have significant activity against S pneumoniae and Chlamydia, whereas ciprofloxacin is more useful against P aeruginosa. Fluoroquinolones have several favorable pharmacologic properties in addition to their broad spectrum. These include oral administration, low protein binding, good distribution to all body compartments, penetration of phagocytes, and a prolonged serum half-life that allows once- or twice-a-day dosing. Levofloxacin and ciprofloxacin are excreted primarily by the kidney, resulting in high drug concentrations in the urine, making them suitable for the treatment of many urinary tract infections. Moxifloxacin is secreted to a smaller degree into the urine.

Because of their broad spectrum and oral administration, fluoroquinolones have been prescribed heavily for many years. This has both increased bacterial resistance and unmasked concerning potential side effects, including tendon injury, diarrhea, cardiac arrhythmias, and peripheral neuropathy. For these reasons, fluoroquinolones are no longer first-line treatment for common problems such as urinary tract infections or bacterial sinusitis.

Folate Inhibitors

Agents that interfere with the synthesis of folic acid by bacteria have selective toxicity because mammalian cells are unable to accomplish this feat, instead using preformed folate from dietary sources. Folic acid is derived from para-aminobenzoic acid (PABA), glutamate, and a pteridine unit. In its reduced form, it is an essential coenzyme for the transport of one-carbon compounds in the synthesis of purines, thymidine, and some amino acids; thus, folic acid is indirectly essential for the synthesis of nucleic acids and proteins. The major inhibitors of the folate pathway are the sulfonamides, trimethoprim, para-aminosalicylic acid, and the sulfones.

Sulfonamides

Sulfonamides are structural analogs of PABA and compete with it for the enzyme (dihydropteroate synthetase) that combines PABA and pteridine in the initial stage of folate synthesis. This blockage has multiple effects on the bacterial cells; the most important of these is disruption of nucleic acid synthesis. The effect is bacteriostatic, and the addition of PABA to a medium that contains sulfonamide neutralizes the inhibitory effect and allows growth to resume.

When introduced in the 1940s, sulfonamides had a very broad spectrum, but resistance developed quickly. Now their primary use is for uncomplicated urinary tract infections caused by members of the Enterobacteriaceae, particularly Escherichia coli. Sulfonamides are convenient for this purpose because they are inexpensive, well absorbed by the oral route, and excreted in high levels in the urine. They also have a role in some skin infections due to MRSA.

Trimethoprim-Sulfamethoxazole

Trimethoprim acts on the folate synthesis pathway but at a point after sulfonamides. It competitively inhibits the activity of bacterial dihydrofolate reductase, which catalyzes the conversion of folate to its reduced active coenzyme form. When combined with sulfamethoxazole, a sulfonamide, trimethoprim leads to a two-stage blockade of the folate pathway, which often results in synergistic bacteriostatic or bactericidal effects. This quality is exploited in therapeutic preparations that combine both agents in a fixed proportion designed to yield optimum synergy.

Trimethoprim-sulfamethoxazole (TMP-SMX) has a spectrum that is much broader and more stable than either of its components alone; this includes most of the common pathogens, whether they are gram-positive or gram-negative, cocci or bacilli. Anaerobes and P aeruginosa, however, are not covered. It is also active against some uncommon agents such as Nocardia. TMP-SMX is widely and effectively used in the treatment of urinary tract infections, otitis media, sinusitis, prostatitis, and MRSA skin infections. Interestingly, its spectrum expends beyond bacteria. It is useful for the treatment of certain protozoan causes of diarrhea, and is the agent of choice for pneumonia caused by Pneumocystis jirovecii, a fungus.

Metronidazole

Metronidazole is a nitroimidazole, a family of compounds with activity against bacteria, fungi, and parasites. The antibacterial action requires reduction of the nitro group under anaerobic conditions, which explains the limitation of its activity to bacteria that prefer anaerobic or at least microaerophilic growth conditions. The reduction products act on the cell at multiple points; the most lethal of these effects is induction of breaks in DNA strands.

Metronidazole is active against a wide range of anaerobes, including B fragilis. Clinically, it is useful for any infection in which anaerobes may be involved, especially those in the gastrointestinal tract. Because these infections are typically polymicrobial, a second antimicrobial (eg, β-lactam) is usually added to cover aerobic and facultative bacteria. Toxicity includes nausea, a metallic taste perversion, and—less commonly—peripheral neuropathy. Alcohol consumption may trigger an unpleasant reaction for the patient.

Rifamicins

Rifampin binds to the β-subunit of DNA-dependent RNA polymerase, which prevents the initiation of RNA synthesis. This agent is active against most gram-positive bacteria and selected gram-negative organisms, including Neisseria and Haemophilus but not members of the Enterobacteriaceae. The most clinically useful property of rifampin is its antimycobacterial activity, which includes Mycobacterium tuberculosis and the other species that most commonly infect humans. Because resistance by mutation of the polymerase readily occurs, rifampin is combined with other agents in the treatment of active infections. It is only used alone for chemoprophylaxis of N meningitidis and H influenzae in close contacts of infected patients, and in the treatment of latent tuberculosis infection. When given for prolonged courses, rifampin may radically alter the metabolism of other medications via induction of hepatic cytochrome enzyme expression. A related drug, rifaximin, is not absorbed when taken by mouth, making it ideal for the treatment and prevention of certain causes of bacterial diarrhea.

KEY CONCLUSIONS

Antimicrobials Acting on the Outer and Cytoplasmic Membranes

Daptomycin is a lipopeptide antimicrobial. This drug’s molecular structure mimics that of the bacterial outer membrane phospholipid bilayer; it inserts itself into this membrane and forms pores that allow efflux of ions, thus killing the cell. Its spectrum is limited to gram-positive organisms, including multidrug-resistant strains of Enterococcus and S aureus. Surfactant molecules in the lung bind to this molecule, rendering it unreliable for the treatment of pneumonia.

The polypeptide antimicrobial agents polymyxin B and colistin have a cationic detergent-like effect. They bind to the cell membranes of susceptible gram-negative bacteria and alter their permeability, resulting in the loss of essential cytoplasmic components and bacterial death. These agents react to a lesser extent with cell membranes of the host, resulting in nephrotoxicity and neurotoxicity. Their spectrum is essentially gram-negative; they act against P aeruginosa and other gram-negative rods. Although these antimicrobials were used for systemic treatment in the past, their use was subsequently limited to topical applications because of their toxicity; with the rise in gram-negative resistance to first-line drugs, these medications are once again being used more by the intravenous route. They have an advantage: resistance to them rarely develops.

Other Agents

Several other effective antimicrobials are in use almost exclusively for a single infectious agent or types of infections such as tuberculosis, urinary tract infections, and anaerobic infections. Where appropriate, these agents will be discussed in the relevant chapter. It is beyond the scope and intent of this book to provide comprehensive coverage of all available agents.

SUSCEPTIBILITY AND RESISTANCE

Deciding whether any bacterium should be considered susceptible or resistant to an antimicrobial involves an integrated assessment of in vitro activity, pharmacologic characteristics, and clinical factors. Any agent approved for clinical use has demonstrated in vitro its potential to inhibit the growth of some target group of bacteria at concentrations that can be achieved with acceptable risks of toxicity. That is, the minimum inhibitory concentration (MIC) can be comfortably exceeded by doses tolerated by the patient. Use of the antimicrobial in animal models and then human infections must also have demonstrated a therapeutic response. Because the influence of antimicrobials on the natural history of different categories of infection (eg, pneumonia, meningitis, diarrhea) varies, clinical trials may include both a range of bacterial species and different infected sites (eg, lung, bone, CSF). These clinical studies are important to determine whether what should work actually does work and, if so, to define the parameters of success and failure. Physicians must decide whether the evidence used to earn FDA approval for a new antimicrobial applies to the particular circumstances of the case before them.

Once these factors are established, the routine selection of therapy can be based on known or expected characteristics of organisms and pharmacologic features of antimicrobial agents. With regard to organisms, use of the term susceptible (sensitive) implies that their MIC is at a concentration attainable in the blood or other appropriate body fluid (eg, urine) using recommended doses. Resistant, the converse of susceptible, implies that the MIC is not exceeded by normally attainable levels. As in all biological systems, the MIC of some organisms lies in between the susceptible and resistant levels. Borderline strains are called intermediately sensitive. The antimicrobial in question may still be used to treat these organisms, but at increased doses. For example, less toxic antibiotics such as the penicillins and cephalosporins can be administered in massive amounts and may thereby inhibit some pathogens that would normally be considered resistant in vitro. Furthermore, in urinary infections, urine levels of some antimicrobial agents may be very high (eg, fluoroquinolones), and organisms that are resistant in vitro may be eliminated in the patient.

Important pharmacologic characteristics of antimicrobial agents include dosage as well as the routes and frequency of administration. Other characteristics include whether the agents are absorbed from the upper gastrointestinal tract, whether they are excreted and concentrated in active form in the urine, whether they can pass into cells, whether and how rapidly they are metabolized, and the duration of effective antimicrobial levels in blood and tissues. Most agents are bound to some extent to serum albumin, and the protein-bound form is usually unavailable for antimicrobial action. The amount of free to bound antibiotic can be expressed as an equilibrium constant, which varies for different antibiotics. In general, high degrees of binding lead to more prolonged but lower serum levels of an active antimicrobial after a single dose.

LABORATORY TESTING OF ANTIMICROBIAL SUSCEPTIBILITY

A unique feature of laboratory testing in bacteriology is that the individual patient’s isolate is routinely tested against a battery of antimicrobial agents. These tests are built around the common theme of placing the organism in the presence of varying concentrations of the antimicrobial in order to determine the MIC. The methods used are standardized, including a measured inoculum of the bacteria and controlled growth conditions (eg, medium, temperature, atmosphere, and time).

In selecting therapy, clinicians must consider more than the results of laboratory tests. The clinical pharmacology of the drug, the cause of the disease, the site of infection, the immune function of the patient, and the pathology of the lesion must be taken into account as well. For example, the antimicrobial must reach the subarachnoid space and cerebrospinal fluid in meningitis. Similarly, treatment may be ineffective for an infection that has resulted in abscess formation unless the abscess is surgically drained. Previous clinical experience is also critical. In typhoid fever, for instance, azithromycin may be effective while aminoglycosides are not, even though the typhoid bacillus may be equally susceptible to both in vitro. This is due to the aminoglycosides’ failure to achieve adequate concentrations inside the macrophages where Salmonella enterica serovar Typhi multiplies.

Dilution Tests

Dilution tests determine the MIC directly by using serial dilutions of the antimicrobial agent in broth that span a clinically significant range of concentrations. The dilutions are prepared in tubes or microdilution wells, and by convention, their concentrations are doubled using a base of 1 μg/mL (0.25, 0.5, 1, 2, 4, 8, and so on). The bacterial inoculum of the patient’s isolate is adjusted to a standard (10⁵ to 10⁶ bacteria/mL) and added to the broth. After incubation overnight (or other defined time), the tubes are examined for turbidity produced by bacterial growth. The first tube in which visible growth is absent (clear) is the MIC for that organism (Figure 23–5).

Automated Tests

Instruments are now available that carry out rapid, automated variants of the broth dilution test. In these systems the bacteria are incubated with the antimicrobial in specialized modules that are read automatically on a frequent basis. The multiple readings and the increased sensitivity of determining endpoints by turbidimetric or fluorometric analysis makes it possible to generate MICs in as little as 4 hours. In laboratories with sufficient volume, these methods are no more expensive than manual methods, and the rapid results have enhanced potential to influence clinical outcome, particularly when interfaced with computerized hospital information systems.

Diffusion Tests

In diffusion testing (often called the Kirby-Bauer technique), the inoculum is seeded onto the surface of an agar plate, and filter paper disks containing defined amounts of antimicrobials are applied. While the plates are incubating, the antimicrobial diffuses into the medium to produce a circular gradient around the disk. After incubation overnight, the size of the zone of growth inhibition around the disk (Figure 23–6A) can be used as an indirect measure of the MIC of the organism. It is also influenced by the growth rate of the organism, the diffusibility of the drug, and other technical factors. The diameters of the zones of inhibition obtained with the various antibiotics are converted to “susceptible,” “intermediate,” or “resistant” categories by referring to a table. This method is convenient and flexible for rapidly growing aerobic and facultative bacteria such as the Enterobacteriaceae, Pseudomonas, and staphylococci. Another diffusion procedure uses gradient strips to produce elliptical zones that can be directly correlated with the MIC. This method, the epsilometer test (Figure 23–6B), can also be applied to slow-growing, fastidious, and anaerobic bacteria. This approach is slower and more laborious than automated broth systems, but it has the advantage of revealing the presence of multiple colony morphologies, mixed infections, or resistant subpopulations that appear as “inner colonies” within an otherwise clear zone of inhibition.

Molecular Testing

The molecular techniques of nucleic acid hybridization, sequencing, and amplification (see Chapter 4) have been applied to the detection and study of resistance. The strategy is to detect the resistance gene rather than to measure the phenotypic expression of that gene’s product. These methods offer the prospect of automation and rapid results, but they can only detect genes already known to science, although some forms of resistance do not yet have well-defined genetic causes. Phenotypic gene expression remains the “bottom line” that guides most resistance testing today.

Bactericidal Testing

The above methods do not distinguish between inhibitory and bactericidal activity. To do so requires quantitative subculture of the clear tubes in the broth dilution test and comparison of the number of viable bacteria at the beginning and end of the test. The least amount required to kill a predetermined portion of the inoculum (usually 99.9%) is called the minimal bactericidal concentration (MBC). Direct bactericidal testing is important in the initial characterization and clinical evaluation of antimicrobial agents but is rarely needed in individual cases. Most of the antimicrobials used for acute and life-threatening infections (eg, β-lactams, aminoglycosides) act by bactericidal mechanisms.

Antimicrobial Assays

For antimicrobials with toxicity near the therapeutic range, monitoring the concentration in the serum or other body fluid is sometimes necessary. Therapeutic monitoring may also be required when the patient’s pharmacologic handling of the agent is unpredictable, as in renal failure. A variety of biologic, immunoassay, and chemical procedures have been developed for this purpose. The drugs most commonly measured are vancomycin and the aminoglycosides.

BACTERIAL RESISTANCE TO ANTIMICROBIALS

Antimicrobial agents were originally hailed as “wonder drugs.” Unfortunately, their effectiveness has been steadily eroded by the appearance of resistant bacterial strains. This resistance may be inherent to the organism or appear in a previously susceptible species by mutation or the acquisition of new genes. Keeping ahead of the microbes requires that we understand the mechanisms by which bacteria develop resistance, and the ways this resistance spreads. The following sections discuss the biochemical mechanisms of resistance, how resistance is genetically controlled, and how resistant strains survive and spread in our society. How these features relate to the antimicrobial groups is summarized in Table 23–2 and further discussed in the chapters on specific bacteria (see Chapters 24-41).

Antimicrobial resistance has survival value for the organism, and its expression in the medical setting requires that virulence be retained despite the change that mediates resistance. There are no direct connections between resistance and virulence: some highly drug-resistant bacteria may cause relatively indolent infections, whereas exquisitely susceptible organisms may cause severe infections. Resistant bacteria have increased opportunities to produce disease, but the disease itself is the same as that produced by the bacterium’s susceptible counterpart. Although uncommon, it is possible for enhanced virulence traits to be added to resistant strains by linkage with virulence genes on plasmids or other genetic elements. This appears to have occurred with the emergence of MRSA clones with enhanced potential to infect skin and soft tissues (see Chapter 24). The term “superbug,” increasingly used to describe multiresistant bacteria, implies this linkage is more common than it actually is.

Mechanisms of Resistance

The major mechanisms of bacterial resistance (Figure 23–7) are (1) Exclusion of the antimicrobial from the bacterial cell due to impermeability or active efflux; (2) alterations of an antimicrobial target, which render it insusceptible; and (3) inactivation of the antimicrobial agent by an enzyme produced by the microorganism.

Exclusion

An effective antimicrobial must enter the bacterial cell and achieve concentrations sufficient to act on its target (Figure 23–8). The cell wall, particularly the outer membrane, of gram-negative bacteria presents a formidable barrier for access to the interior of the cell. Inability to traverse the outer membrane is the primary reason most β-lactams are less active against gram-negative than gram-positive bacteria. Outer membrane protein porin channels may allow drug penetration depending on their size, charge, degree of hydrophobicity, or general molecular configuration. This is a major reason for inherent resistance to antimicrobial agents, but these transport characteristics may change even in typically susceptible species due to mutations in the porin proteins. For example, strains of P aeruginosa commonly develop resistance to carbapenems due to loss of the outer membrane protein most important for their penetration.

Some antimicrobials must be actively transported into the cell. For example, bacteria which lack the metabolic pathways required for transport of aminoglycosides across the cytoplasmic membrane (streptococci, enterococci, anaerobes) are intrinsically resistant. Conversely, other antimicrobials are actively transported out of the cell. A number of bacterial species have energy-dependent efflux mechanisms that literally pump antimicrobial agents which have entered the cell back out. The membrane transporter systems which drive these efflux pumps often affect antimicrobials of several classes.

Altered Target

Once in the cell, antimicrobials act by binding and inactivating their target, which is typically a crucial enzyme or ribosomal site (Figure 23–9). If the target is altered in a way that decreases its affinity for the antimicrobial, the inhibitory effect will be proportionately decreased. Substitution of a single amino acid at a certain location in a protein may alter its binding to the antimicrobial without affecting its function in the bacterial cell.

If an alteration at a single site on the target renders it nonsusceptible, mutation to resistance can occur in a single step, even during therapy. This occurred with the early aminoglycosides (streptomycin), which bound to a single ribosomal site, and the first quinolone (nalidixic acid), which attached to only one of four possible topoisomerase subunits. Newer agents in each class bind at multiple sites on their target, making mutation to resistance less probable.

One of the most important examples of altered target involves the β-lactam family and the peptidoglycan transpeptidase penicillin-binding proteins (PBPs) on which they act. In widely divergent gram-positive and gram-negative species, changes in one or more of these proteins correlate with decreased susceptibility to multiple β-lactams. These alterations were initially detected as changes in electrophoretic migration of one or more PBPs using radiolabeled penicillin (hence the origin of the term PBP). These changes have now been traced to point mutations, substitutions of amino acid sequences, and even synthesis of a new enzyme.

Because the altered binding is not absolute, decreases in susceptibility are incremental and often small. Wild-type pneumococci and gonococci are inhibited by 0.06 μg/mL of penicillin, while those with altered PBPs have MICs of 0.1 to 8.0 μg/mL. At the lower end, these MICs still appear to be within therapeutic range but are associated with treatment failures, even when dosage is increased. Altered PBPs may affect some or all β-lactams. Although the exact MICs vary, a strain with a 10-fold decrease in susceptibility to penicillin has decreased susceptibility to cephalosporins to about the same degree.

PBP alterations are the prime reason for emergence of penicillin-resistant pneumococci and MRSA. They are one of multiple mechanisms of resistance in a variety of other bacteria including enterococci, gonococci, H influenzae, and many other gram-positive and gram-negative species.

Alteration of the target does not require mutation and can occur by the action of a new enzyme produced by the bacteria. Vancomycin-resistant enterococci have enzyme systems that substitute a different amino acid in the terminal position of the peptidoglycan side chain (often alanyl lactate instead of alanyl alanine). Vancomycin does not bind to the alternate amino acid, rendering these strains resistant. Resistance to sulfonamides and trimethoprim occurs by acquisition of new enzymes with low affinity for these agents but still allows bacterial cells to carry out their respective functions in the folate synthesis pathway.

Clindamycin resistance involves an enzyme that methylates ribosomal RNA, preventing attachment. This modification also confers resistance to erythromycin and other macrolides, because they share binding sites. Interestingly, induction with erythromycin leads to clindamycin resistance, although the reverse is unusual.

Enzymatic Inactivation

Enzymatic inactivation of antimicrobial agents is the most powerful and robust resistance mechanism (Figure 23–10). Literally hundreds of distinct enzymes produced by resistant bacteria may inactivate antimicrobials in the cell, in the periplasmic space, or outside the cell. They may act on the antimicrobial molecule by disrupting its structure or by catalyzing a reaction that chemically modifies it.

β-Lactamases

β-Lactamase is a general term referring to any one of many bacterial enzymes able to break open the β-lactam ring and inactivate various members of the β-lactam group. The first was discovered when penicillin-resistant strains of S aureus emerged and were found to inactivate penicillin in vitro. The enzyme was called penicillinase, but with expansion of the β-lactam family and concomitant resistance, it has become clear that the situation is quite complex. Each β-lactamase is a distinct enzyme with its own physical characteristics and substrate profile. For example, the original staphylococcal penicillinase is also active against ampicillin but not against methicillin or any cephalosporin. β-Lactamases produced by E coli may have some cephalosporinase activity but vary in their potency against individual first-, second-, third-, and fourth-generation cephalosporins. Some β-lactamases are bound by the β-lactamase inhibitor clavulanic acid, while others are not.

Bacteria that produce β-lactamases typically demonstrate high-level resistance with MICs far outside the therapeutic range. But even weak β-lactamase producers are considered resistant because the outcome of susceptibility tests (and presumably infected sites) is strongly influenced by the number of bacteria present. Large bacterial populations may secrete enough β-lactamase to inactivate the antimicrobial before it even reaches the organisms.

A full discussion of β-lactamase classification is beyond the scope of this book, but some understanding of the major types is useful.

• Most gram-positive β-lactamases are exoenzymes with little activity against cephalosporins or the antistaphylococcal penicillins (methicillin, oxacillin). They are bound by β-lactamase inhibitors such as clavulanic acid.

• Some gram-positive β-lactamases, such as Type A β-lactamase, selectively hydrolyze cefazolin and cephalexin, while remaining ineffective against the antistaphylococcal penicillins.

• Gram-negative enzymes act in the periplasmic space and may have penicillinase and/or cephalosporinase activity. They may or may not be inhibited by clavulanic acid. Many of the gram-negative β-lactamases are constitutively produced at very low levels but can be induced to high level expression by exposure to a β-lactam agent. The resistance gene ampC is a notorious member of this group. AmpC is concerning because its expression may not be induced during routine laboratory testing, but may subsequently be induced and lead to clinical failure during treatment with penicillins or first- and third-generation cephalosporins.

• Even more worrisome is another class of gram-negative resistance genes, called extended-spectrum β-lactamases (ESBLs) because their substrates include multiple cephalosporins. The laboratory detection of ESBLs is complex, as is their naming scheme (CTX-M, TEM, OXA, SHV, etc). From a clinical perspective, they are significant because treatment with any generation of cephalosporin may lead to clinical failure. The newer β-lactamase inhibitor avibactam was designed with these enzymes in mind. Carbapenems are an excellent drug class for treating infections caused by ESBL-producing organisms.

• Most concerning of all among the gram-negative resistance genes are the carbapenemases. Although carbapenems still provide reliable coverage of Enterobacteriaceae in most circumstances, enzymes which specialize in hydrolyzing these drugs—and usually penicillins and cephalosporins at the same time—are on the rise. New Delhi metallo-betalactamase (NDM-1) and Klebsialla pneumoniae carbapenemase (KPC) are but two troubling members of a larger family of such genes. These genes have made carbapenem-resistant Enterobacteriaceae (CRE) one of the most important challenges facing infectious diseases medicine today.

Modifying Enzymes

The most common cause of acquired bacterial resistance to aminoglycosides is through the production of one or more of over 50 enzymes that acetylate, adenylate, or phosphorylate hydroxyl or amino groups on the aminoglycoside molecule. The modifications take place in the cytosol or in close association with the cytoplasmic membrane. The resistance conveyed by these actions is usually high level; the chemically modified aminoglycoside no longer binds to the ribosome. As with the β-lactamases, the aminoglycoside-modifying enzymes represent a large and diverse group of bacterial proteins, each with its characteristic properties and substrate profile. Inactivating enzymes have been described for a number of other antimicrobials. Most act by chemically modifying the antimicrobial molecule in a manner similar to the aminoglycoside-modifying enzymes. The most clinically significant enzymes convey resistance to erythromycin (esterase, phosphotransferase) and chloramphenicol (acetyltransferase).

Genetics of Resistance

Intrinsic Resistance

For any antimicrobial, there are bacterial species that are typically within its spectrum and those which are not (see Appendix 23–1). The resistance of the latter group is referred to as intrinsic or chromosomal to reflect its inherent nature. The resistant species have features such as permeability barriers, a lack of susceptibility of the cell wall, or ribosomal targets that make them inherently insusceptible. Some species constitutively produce low levels of inactivating enzymes, particularly the β-lactamases of gram-negative bacteria. The chromosomal genes encoding these β-lactamases may be under repressor control and subject to induction by certain β-lactam antimicrobials. This leads to increased production of β-lactamase, which usually results in resistance not only to the inducer but other β-lactams to which the organism would otherwise be susceptible. AmpC β-lactamases operate in this manner.

Acquired Resistance

A species may initially be susceptible to an antibiotic, but subsequently develop resistance. Such acquired resistance may be due to a genetic mutation within that organism, or may be derived from another organism by the acquisition of new genes.

Mutational Resistance

Acquired resistance may occur when there is a crucial mutation in the target of the antimicrobial or in proteins related to access to the target (ie, permeability). Mutations in regulatory proteins can also lead to resistance. Mutations take place at a regular but low frequency and are expressed only if they are not associated with other effects that are disadvantageous to the bacterial cell. Mutational resistance can emerge in a single step or evolve slowly, requiring multiple mutations before clinically significant resistance is achieved. Single-step mutational resistance is most likely when the antimicrobial agent binds to a single site on its target. Resistance can also emerge rapidly when it is related to gene regulation, such as mutational derepression of a chromosomally encoded cephalosporinase. A slow, progressive resistance evolving over years, even decades, is typical for β-lactam resistance related to altered PBPs.

Genetic Exchange

Of the four major mechanisms of genetic exchange among bacteria described in Chapter 21 and illustrated in Figure 23–11 (transformation, transduction, conjugation, transposition), conjugation and transposition are the most important clinically and often work in tandem.

Plasmids and Conjugation

The transfer of plasmids by conjugation was the first discovered mechanism for the acquisition of new resistance genes, and it continues to be the most important. Resistance genes on plasmids (R plasmids) can determine resistance to one antimicrobial or to several that act by different mechanisms. After conjugation, the resistance genes may remain on a recircularized plasmid or, less often, become integrated into the chromosome by recombination. A single cell may contain more than one distinct plasmid and/or multiple copies of the same plasmid. Although most resistance mechanisms have been linked to plasmids in one species or another, plasmid distribution among the bacterial pathogens is by no means uniform. The compatibility systems that maintain plasmids from one bacterial cell generation to the next are complex. Some species of bacteria are more likely than others to contain plasmids. For example, Neisseria gonorrhoeae typically has multiple plasmids, whereas closely related N meningitidis rarely has any.

Plasmids are most likely to be transferred to another strain if they are conjugative, that is, if the resistance plasmid also contains the genes mediating conjugation. Another factor in the spread of plasmids is their host range. Some plasmids can be transferred only to closely related strains; others can be transferred to a broad range of species within and beyond their own genus. A conjugative plasmid with a broad host range has great potential to spread any resistance genes it carries.

Transposons and Transposition

Transposons containing resistance genes can move from plasmid to plasmid or between plasmid and chromosome. Most of the resistance genes carried on plasmids are transposon insertions that can be carried along with the rest of the plasmid genome to another strain by conjugation. Once there, the transposon is free to remain in the original plasmid, insert into a new plasmid, insert into the chromosome, or any combination of these (Figure 23–11). Theoretically, plasmids can accomplish the same events by recombination, but the nature of the transposition process is such that it is much more likely to result in the transfer of an intact gene. Transposons also have a variable host range which in general is even broader than plasmids. Together, conjugation and transposition provide extremely efficient means for spreading resistance genes.

Other Genetic Mechanisms

Transduction is the process in which viral bacteriophages inject genetic material into bacteria. Although the transfer of resistance genes by transduction has been demonstrated in the laboratory, its association with clinically significant resistance has been uncommon. Transduction of imipenem resistance by wild-type bacteriophages carried by P aeruginosa to other strains of the same bacteria is one such example. Because of the high specificity of bacteriophages, transduction is typically limited to bacteria of the same species. Transformation is the insertion of DNA directly across the cell membrane. This is the most common way genes are manipulated in the laboratory, but detecting its occurrence in the environment or human hosts is particularly difficult, because naked DNA lacks the signatures that flag the presence of plasmids and transposons. Molecular epidemiologic studies suggest that the spread of PBP mutations in S pneumoniae is due to transformation, and there may be more examples awaiting discovery.

Epidemiology of Resistance

It seems that sooner or later microorganisms will develop resistance to any antimicrobial agent to which they are exposed. Since the start of the antibiotic era, each new antimicrobial has tended to go through a remarkably similar sequence. When an agent is first introduced, its spectrum of activity is highly predictable; some species are naturally resistant, and others are susceptible, with few exceptions. With clinical use, resistant strains of previously susceptible species begin to appear and become increasingly common.

In some situations, resistance develops rapidly; in other cases it takes years, or even decades. For example, when penicillin was first introduced in 1944, all strains of S aureus appeared to be fully susceptible, but by 1950, less than one-third of isolates remained susceptible. We now know that strains containing the penicillinase plasmid existed long before and were selected when penicillin use became widespread. These plasmids likely conferred a survival benefit to strains of S aureus in the environment, where they live in competition with Penicillium and other molds. However, the discovery of H influenzae (meningitis) and N gonorrhoeae (gonorrhea) strains resistant to ampicillin and penicillin did not occur until those antibiotics had been used heavily for a decade or more. In these instances, resistance genes apparently not present in the species initially were acquired from other bacterial species, either directly or through recombination of plasmids. There are small enclaves of bacteria that have not developed resistance. After almost a century, the causes of syphilis (Treponema pallidum) and strep throat (group A streptococcus) have thus far retained their susceptibility to penicillin.

Whatever the genetic mechanism, the persistence and spread of resistance requires an environment in which the resistant strain pays no price in terms of fitness, or has a selective advantage. The primary human factors which favor this selection are the overuse of antimicrobial agents in medicine and the inclusion of antimicrobials in livestock feeds. Any use of antimicrobial agents by physicians—whether appropriate or not—has the potential for the unintended consequence of selecting for resistance. This includes prescribing antibacterial agents for viral infections, or using a broad-spectrum agent when a narrower drug would work just as well—if not better. Exceeding guidelines for prophylactic use of antimicrobials (see later) also contributes. In many nations, physician prescriptions are not required for the use of antibiotics, and self-prescription of antibiotics is common, which may accelerate resistance. Microorganisms do not respect geopolitical boundaries, and the effect of antibiotic misuse in one region can have profound impacts thousands of miles away. As with any intervention in medicine, the use of antimicrobial agents carries benefits and risks for the patient. The difference with antimicrobials is that the risk of resistance is for the population at large, not just the individual patient.

The addition of antimicrobials to animal feeds for their growth-promoting effects is a concerning source of resistant strains of bacteria. Cattle or poultry that consume feed supplemented with antimicrobials rapidly develop resistant enteric flora that spreads throughout the herd. Resistant strains can then appear in the flora of humans living in proximity or handling animal products, including consumers at home. Links from farm to human disease have been established in multiple outbreaks. As a consequence, some countries have banned the use of antimicrobial agents which are used in humans unless necessary for the treatment or prevention of livestock infections. In the United States, a similar movement is now in process.

KEY CONCLUSIONS

ANTIMICROBIAL STEWARDSHIP

Ultimately, bacteria will always evolve in response to selective pressure. Because this has the potential to happen much more rapidly than we can develop new antibiotics, we must defend the current armamentarium. In the same way that we must protect our natural environment, so too should we use this precious resource wisely.

A coordinated, sustained effort will be required to minimize the spread of antimicrobial resistance. Everyone shares responsibility for the current crisis of resistance—the veterinarians who use it in farm animals, the politicians who regulate and set priorities in healthcare, the insurance companies that dictate access to certain medications for reasons of cost, the drug manufacturers who choose which drugs to focus on, the patients who ask for antibiotics even when they are not necessary, and of course the providers who prescribe these vital medications. The coordinated response to antibiotic resistance is called “antimicrobial stewardship.” Many hospitals now have formal stewardship programs, closely integrated with infection prevention teams. Most antibiotics are prescribed in the outpatient setting where there is a pressing need for better stewardship. As a future prescriber, you bear a professional responsibility to become an antimicrobial steward for the benefit of the individual patient, and for the benefit of society. The mantra is “Together, we can reduce antimicrobial resistance.”

Learning to prescribe antibiotics effectively and safely takes practice. Some fundamental principles are included in the following discussion, and in Appendix 23–2.

Empiric Therapy

Unfortunately, definitive microbiological data is rarely available when patients first present with an infection. Because time is usually of the essence, providers must make their best guess and start with “empiric therapy.” These first decisions are based on the physician’s assessment of the probable microbial etiology of the patient’s infection. Variables involved in choosing the best empiric drug include the site of infection (eg, throat, lung, urine, bone) and epidemiologic factors such as season, geography, patient age, pregnancy status, drug allergies, prior antibiotic exposure, other medications being taken, and predisposing conditions. This list of individual factors must then be matched with their probable microbiology and antimicrobial susceptibilities as shown in Table 23–1 and Appendix 23–1. Local antibiograms provide “batting averages” for each antimicrobial against common bacterial pathogens. These are available from hospital laboratories and infection control committees; note that, depending on the technique used to create the antibiogram, these resources may be more suitable for inpatients than outpatients, because resistance to broad-spectrum agents may be less rampant in the community than in the hospital.

This process may be as simple as selecting penicillin to treat an ambulatory patient with suspected group A streptococcal pharyngitis, or as complex as resorting to a cocktail of broad-spectrum antibacterial, antifungal, and antiviral agents to treat a critically ill inpatient who has undergone stem cell transplantation. In general, the risks of broad-spectrum treatment (eg, drug toxicity and selection of resistant flora) become more acceptable as the severity of the infection increases. When the risk of not “covering” an improbable pathogen is death, as may be the case in critically ill or immunosuppressed patients, it is more difficult to prescribe narrow coverage initially. Empiric therapy should be converted to specific therapy within a few days, once microbiology data are available, although in some instances this is not possible. For example, in otitis media there is no easy way to culture the middle ear so empiric therapy must be continued and the outcome evaluated on clinical grounds.

Specific Therapy

Specific therapy is that directed only at the known pathogen, based on isolation and susceptibility testing of the patient’s organism in the laboratory. This is possible for almost all bacterial infections—if microbiological testing is performed. As the results of Gram stains, cultures, and susceptibility tests are reported, unnecessary antimicrobials must be discontinued and the spectrum of therapy narrowed. For example, a patient with suspected staphylococcal or streptococcal infection might be empirically started on vancomycin, which covers both possibilities. Once MRSA has been excluded, a more specific β-lactam is substituted for the broader spectrum treatment. Usually this is a single best agent, but sometimes combinations of antimicrobials that have different modes of action are used for enhanced effect. The major indications for combinations are to reduce the probability of emergence of resistance (which is important in chronic infections like tuberculosis and lung infections in cystic fibrosis), and taking advantage of known synergy between two antimicrobials. Synergy happens when the activity of a drug combination is far greater than would be expected from the individual MICs of the two antimicrobials.

Prophylaxis

The use of antimicrobials to prevent infection is a tempting but potentially hazardous endeavor. The risk for the individual patient: subsequent infection with a different, more resistant organism. The risk for the whole population: increased risk for the spread of resistance. After many years of experience, the indications for antimicrobial prophylaxis have been narrowed to a small number of situations in which antimicrobials have been shown to decrease transmission during a period of high risk. For example, persons known to have been exposed to highly infectious and virulent pathogens like N meningitidis (meningitis), Bacillus anthracis (anthrax), or Yersinia pestis (plague) can abort an infection during the incubation period by the administration of ciprofloxacin. Prophylaxis can also reduce the risk of endogenous infection associated with certain surgical and dental procedures if given during the procedure. The practice of administering prophylactic penicillin during labor to mothers with demonstrated vaginal group B streptococcal (GBS) colonization dramatically decreases the leading cause of sepsis and meningitis in neonates.