Principles of Anti-Infective Therapy

Selecting Antimicrobial Therapy

Empirical Antimicrobial Therapy

It is rare that antimicrobial therapy is instituted for a serious infection with full knowledge of the identity and antibiotic susceptibilities of the specific pathogen, so empirical antibiotic therapy typically must be initiated based on presumptions regarding the nature of the infection. The first challenge is to assess whether the patient likely has an infection and, if so, where is it located. The patient’s clinical manifestations or syndromes may not be infectious. Autoimmune diseases ( Chapter 236 ), lymphomas ( Chapters 171 and 172 ), drug reactions ( Chapter 234 ), hematomas, and deep vein thromboses ( Chapter 68 ) often present with fever as a cardinal symptom, so it is important to keep in mind that empirical antibiotics may not be needed. In other situations, a careful history, thorough physical examination, and a few laboratory tests will allow identification of a clinical syndrome, such as meningitis ( Chapter 381 ), pneumonia ( Chapter 85 ), urinary tract infection ( Chapter 263 ), enteric infection ( Chapter 262 ), or sexually transmitted disease ( Chapter 264 ). Some infectious syndromes, such as upper respiratory tract infections ( Chapter 84 ), may be sufficiently likely to be caused by virus such that antibiotics are likely to be of no value.

After the infectious syndrome is identified, the next step is to make a reasonable estimation of the specific pathogens that are likely to be involved. Appropriate empiric therapy is especially important when delayed effective antimicrobial coverage is associated with worse clinical outcomes, such as for community-acquired pneumonia ( Chapter 85 ), central nervous system infections ( Chapters 381 to 383 ), and sepsis ( Chapter 94 ). In an adult patient with suspected bacterial meningitis ( Chapter 381 ), for example, ceftriaxone covers both of the two most likely pathogens, Streptococcus pneumoniae ( Chapter 268 ) and Neisseria meningitidis ( Chapters 274 ). If the patient is elderly, pregnant, or otherwise at high risk if the pathogen is Listeria monocytogenes ( Chapter 272 ), the regimen could be expanded to include ampicillin. If there is a reason to suspect S. pneumoniae with reduced susceptibility to ceftriaxone, vancomycin may be added. Finally, if the clinical picture is equivocal and includes the possibility of encephalitis, acyclovir might be added to address the possibility of herpes simplex encephalitis ( Chapters 345 and 383 ). The difficulty of distinguishing among pathogens when treating community-acquired pneumonia ( Chapter 85 ) leads many to prescribe both ceftriaxone (for S. pneumoniae [ Chapter 268 ] and Haemophilus influenzae [ Chapter 277 ]) and azithromycin (for atypical pathogens Legionella pneumophila [ Chapter 290 ], Chlamydia pneumoniae [ Chapter 294 ], and Mycoplasma pneumoniae [ Chapter 293 ]). Conversely, if a patient presents with a community-acquired pyelonephritis ( Chapter 263 ), Escherichia coli (not S. pneumoniae ) should be at the top of the list. The likely pathogens can then guide what antimicrobial agent(s) to administer.

This syndrome to pathogen to antimicrobial agent thought process is designed to allow identification of an antibiotic regimen that is as targeted as possible, so as to have less impact on the resident flora and a lower likelihood of promoting the emergence of resistance. For example, a typical cellulitis will be most often caused by Streptococcus pyogenes , and less commonly, by Staphylococcus aureus . These two pathogens could be adequately treated by nafcillin (which has only relatively narrow gram-positive coverage) or by ampicillin-sulbactam (which has broad coverage that includes facultative and anaerobic gram-negative bacilli and a wide range of gram-positive cocci). It is obviously preferable to avoid killing the facultative gram-negative and anaerobic bacteria in the gastrointestinal tract, especially when the likelihood that they are responsible for the infection is very low. Conversely, very broad antibacterial coverage may be critical to success in other situations, such as treating an intra-abdominal infection from a perforated appendix ( Chapter 128 ), a situation in which the polymicrobial microbiology of the likely infection will never be precisely delineated.

When considering empirical coverage, it is useful to have as much information as possible. Relevant information might include a Gram stain of material from the site of the infection. When Gram stains are not available or relevant, specific or multiplex polymerase chain reaction (PCR) amplification may give a sense of whether broader coverage is necessary.

The other major consideration in choosing an empirical regimen is the condition of the patient. If the patient is febrile but awake and fully functional, the empirical regimen can be more parsimonious, unless the situation worsens acutely. Conversely, the margin for error is much less in septic shock ( Chapter 94 ), for which broad coverage for all conceivable pathogens is indicated until more information is obtained.

Unfortunately, real-life clinical scenarios can be challenging. For example, the severity of a patient’s illness may be due to noninfectious causes, and few antibiotics are truly narrow-spectrum. As a result, the best strategy to prevent antimicrobial side effects and the emergence of resistance is to minimize exposure to antibiotics by treating only infections likely to respond to antibiotics with the right dose for the shortest effective duration possible.

Definitive Antimicrobial Treatment

Definitive therapy requires the identification of a pathogen (or pathogens) responsible for the infection as well as their susceptibility to specific antimicrobial agents. This task is sometimes relatively straightforward (e.g., multiple blood cultures growing S. aureus in a patient with clinical and imaging evidence of endocarditis) and sometimes more complicated (e.g., growing coagulase-negative staphylococci, S. pyogenes , Pseudomonas aeruginosa , and Bacteroides fragilis from deep cultures of a diabetic foot ulcer). Once the pathogen(s) are identified and the susceptibility to antimicrobial agents is determined, empirical broad-spectrum therapy often can be narrowed to less toxic, less complicated, or less expensive agents. In some circumstances, it may be possible to switch from a parenteral to an oral regimen, thereby reducing the risks associated with an intravenous catheter and perhaps shortening the patient’s duration of hospitalization. Alternatively, culture and susceptibility testing may guide expanded regimens to include coverage for an unanticipated or resistant pathogen.

Whenever a regimen is chosen, the clinician should also consider the potential next alternative antimicrobial therapy if the first choice must be discontinued because of an allergic reaction or drug toxicity prior to completing the therapeutic course. Such contingencies should be considered in advance because they may impact the original choice of antibiotic. For example, if a patient develops a rash while being co-administered both a penicillin and a cephalosporin, both classes of drugs may need to be avoided.

Susceptibility Testing

A cornerstone of the treatment of infectious diseases is in vitro susceptibility to determine the ability of a microorganism to grow in the presence of serial dilutions of a variety of potentially useful antimicrobial agents. In standard susceptibility tests, serial dilutions of a given antibiotic are made in either broth or agar, and then a specified number of bacteria are inoculated into (broth) or onto (agar) the media and incubated at temperatures and other conditions (e.g., aerobic and anaerobic) permissive for growth. Growth is determined by visual inspection after 16 to 24 hours. The lowest antibiotic concentration at which no visible growth is seen is defined as the minimal inhibitory concentration (MIC). An alternative agar-based test is the disc diffusion assay, in which a paper disc containing a fixed concentration of an antibiotic is placed onto an agar plate previously spread with a specified number of bacteria. The antibiotic diffuses into the agar and inhibits growth of the strain at distances with sufficient concentrations of diffused antibiotic. Zone sizes are measured, and susceptibility is inferred. Another alternative is the E-test, in which a strip containing a gradient of the antibiotic is placed on an agar plate that was previously inoculated with the bacteria.

Each licensed antibiotic has a defined minimal inhibitory concentration that classifies an organism as “susceptible” or “resistant” (and sometimes as “intermediate” as well). These determinations are based on obtainable blood and tissue levels in vivo to assist the clinician in understanding the minimal inhibitory concentration number and its therapeutic implications. However, designations of “susceptible” and “resistant” are best understood as general guides, because exceptions can result in an unnecessary limitation of options in some cases and inadequate therapy in others. For example, the minimal inhibitory concentrations of a typical strain of P. aeruginosa ( Chapter 282 ) fall well outside of the susceptible range for tetracycline, but tetracycline is so heavily concentrated in the urine that it achieves levels that exceed even the elevated minimal inhibitory concentrations of P. aeruginosa and can successfully treat P. aeruginosa cystitis. Conversely, pneumonia strains caused by S. pneumoniae with an minimal inhibitory concentration of 2 µg/mL would be readily treatable with intravenous ceftriaxone, but ceftriaxone’s limited penetration into the cerebrospinal fluid (CSF) would mean that strain would be considered resistant to ceftriaxone for treating meningitis. Fine distinctions such as these are best made by expert consultants.

A final point about susceptibility is that it may be dependent on the inoculum of bacteria that are used in the assay. In most cases, a standardized quantity of bacteria is added when performing susceptibility testing. If that inoculum is increased 100-fold, the MIC may increase dramatically, as is often seen with extended-spectrum, β-lactamase–producing Enterobacteriaceae and cephalosporins. As such, it is sometimes important to estimate the inoculum of bacteria as well as the likely penetration of the antibiotic when determining whether an antibiotic that shows in vitro activity will be effective clinically. For example, a dense pneumonia, in which the bacterial inoculum is likely to be very high, may not respond to the same antibiotics that would be quite effective for treating a low-inoculum urinary tract infection caused by the same strain.

Many modern bacteriology laboratories now use automated approaches or genotypic methods, such as PCR testing, to identify bacterial polymorphisms that are indicative of antibiotic resistance. These panels more rapidly provide information when they identify a resistance gene, but they are limited by the fact that they can only find the genes in the panel and not any novel resistance genes. Although use of these panels has not been shown to improve clinical outcomes, they are quite helpful when looking for specific pathogens, such as identifying nasal colonization with methicillin-resistant S. aureus ( Chapter 267 ) or detecting herpes simplex ( Chapter 345 ) as a cause of viral encephalitis ( Chapter 383 ). Matrix-assisted laser desorption ionization–time of flight mass spectrometry is an alternative to PCR detection methods.

Bactericidal Activity

Standard susceptibility assays test the ability of the antibiotic to inhibit further growth of the microorganism but do not indicate whether the strain is killed by the antibiotic. Most antibiotics are characteristically either bacteriostatic (inhibit growth) or bactericidal (inhibit and kill), but the characteristics may vary with the antibiotic-pathogen pair. Bactericidal activity is determined using time-kill assays, which assess the degree to which an inoculum of bacteria is reduced after a 24-hour period; a greater than 99.9% reduction in the inoculum is required to be designated bactericidal. Fortunately, successful treatment of most infections requires only that the pathogen be inhibited, thereby allowing the immune system time to clear the infection, so bactericidal therapy is not crucial for most infections. In some situations, however, such as in endocarditis (with poor penetration of the immune system into the vegetation), in meningitis (a desire for rapid killing to preserve brain function), and often in osteomyelitis (poor penetration and dormant state of bacteria), bactericidal therapy yields a higher degree of success than bacteriostatic therapy.

Targeting Antimicrobial Therapy Appropriate to the Infection and the Patient

Nature of the Infection

The conditions under which antimicrobial agents are tested for activity against bacteria in the laboratory are very different from the conditions of an actual human infection, and it is remarkable that in vitro susceptibility correlates well with clinical efficacy in many clinical infections. In considering the optimal choice of agent(s), an important first determination is whether the antibiotic will reach the location of the infection in concentrations sufficient to inhibit the infecting microorganisms. For example, the fluoroquinolone antimicrobial agent norfloxacin can be used only to treat urinary tract infections because the urine is the only body fluid in which it achieves sufficient concentrations to be effective. Conversely, the fluoroquinolone moxifloxacin is effective for many infections but is not recommended for urinary tract infections because it does not achieve sufficient concentrations in the urine. These determinations are sometimes species dependent. For example, Enterococcus faecalis ( Chapter 270 ) generally has MICs for the fluoroquinolone ciprofloxacin ranging from 1 to 4 µg/mL. Given the concentrations that ciprofloxacin achieves in the urine, it is a reasonable choice for treatment of a urinary tract infection ( Chapter 263 ) caused by a susceptible strain. However, since serum concentrations of ciprofloxacin at peak are generally in the 2 µg/mL range, it is not adequate for treating tissue enterococcal infections.

The local environment can also have an impact. For example, daptomycin is not sufficiently effective in treating pneumonia because it is inactivated when it complexes with surfactant in the alveoli.

Another major consideration when treating a clinical infection is the state of the microorganisms in the infection. Bacteria reduce their rate of growth when present in a large inoculum, and some antibiotics, notably β-lactams, are more effective at killing bacteria when they are multiplying rapidly and hence may be less effective in some high-inoculum infections. Under many conditions, bacteria also form biofilms, which are carbohydrate/DNA/water matrices that impair the action of antibiotics through a variety of mechanisms, including altered metabolic activity of the microorganism and reduced access for the antibiotic. Under certain circumstances (for example, in chronic osteomyelitis; Chapter 251 ), S. aureus strains can form “small colony variants,” which are also metabolically different and respond poorly to many antimicrobials. Finally, there is the phenomenon known as “persisters,” a minute percentage of a large original inoculum that for some reason is not killed by the antibiotic and persists to grow again once the antimicrobial pressure has subsided. Antimicrobial therapies for these altered states are quite limited in their effectiveness, thereby emphasizing the importance of reducing the inoculum of infection by drainage or debridement whenever possible.

Infection of prosthetic material presents a major challenge. Access of antibiotics to the infection site is often limited by both physical barriers and biofilms. Prosthetic joints, prosthetic cardiac valves, intravascular stents, and fistulae can all become infected. When possible, all prosthetic material should be removed to facilitate treatment of the infection. In cases in which removal is not a viable option, prolonged therapy (often with combinations of agents that include the rapidly bactericidal agent rifampin) is required, sometimes followed by indefinite suppression with an oral agent active against the infecting strain. The goal of indefinite therapy is not to cure the infection but rather to keep it localized. Although rifampin alone is effective in some of these circumstances, resistance can be generated by just a single point mutation in the RNA polymerase gene, thereby making the emergence of resistance a high-frequency event. For this reason, rifampin should always be used in combination with other agents, although the risk of resistance remains even when used in combination.

Host Factors That Impact Choice or Amount of Antibiotic Prescribed

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here