Principles of antibiotic use

Published on 27/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2872 times

Chapter 64 Principles of antibiotic use

The intensive care unit (ICU) is always the area of any hospital associated with the greatest use of antibiotics. Much of this high usage is unavoidable, but the clinician working in the ICU must realise that there is an essential consequence of this use. Antibiotic use which should eliminate susceptible organisms promotes (over)growth of other, non-susceptible organisms, especially fungi. As far as bacteria are concerned, antibiotics confer enormous selective advantage to resistant strains, and therefore these strains will congregate where their advantage is greatest, in the ICU. Resistance (and fungal overgrowth) is a direct consequence of usage, and every course of inappropriate antibiotics should be avoided to help reduce the burden of resistance.

Antibiotic stewardship ¹ has been suggested as a new strategy to help limit resistance. This involves selecting an appropriate drug and optimising its dose and duration to cure an infection while minimising toxicity and conditions for selection of resistant bacterial strains. Inadequate doses of even the ‘correct’ antibiotic may lead to survival of initially susceptible organisms.^2,³ For the optimal use of antibiotics not only should antibiotic pharmacokinetics be understood, but there should be clear and rational principles on which each specific antibiotic prescription in the ICU is based. Also, it is probably better to have portions of the ICU population receive different classes of antibiotics at the same time.

Although this chapter will provide basic principles for most of the antibiotic classes commonly used in the ICU, some important antimicrobial agents will not be specifically addressed here, namely macrolides, clindamycin and the antifungal agents.

GENERAL PRINCIPLES ⁴^–⁸

1 All appropriate microbiological specimens, including blood cultures, should be obtained before commencing antibiotic therapy. An immediate Gram-stained report may indicate the appropriate antibiotic to use; otherwise a ‘best-guess’ choice is made, which is dependent on the clinical situation. This important and common clinical phenomenon involves trying to predict the infecting organism(s).

2 Blood cultures should be taken from a venepuncture site, after adequate skin antisepsis, and not from an intravenous or arterial catheter. Two separate sets of 20 ml (for adults) should be taken, the timing of which is less important.⁹ Depending on what system is used, probably 10 ml should be placed into two different blood culture bottles.

3 Once a decision is made to administer antibiotics, they should be administered without delay.

4 The decision for empiric therapy, i.e. cover for the most ‘likely’ organisms causing any specific infection, must include various factors such as the site of the infecting organism (respiratory tract pathogens differ from those of abdominal infections), community versus hospital-associated infection; recent previous antibiotic prescription; ward versus ICU-acquired infection and knowledge of the organisms commonly grown in patients in any specific area. This latter point is where ward/unit surveillance becomes important.^10,¹¹

5 Whilst there should be an attempt to use a narrow-spectrum antibiotic whenever practicable, appropriate therapy, particularly for empirical choice for nosocomial sepsis, mandates initial broad-spectrum antibiotics, even a combination, until culture results are back,¹² at which time de-escalation should be embarked upon (see below). Inappropriate and/or delayed correct antibiotic use in the ICU has been shown to have an impact on morbidity and mortality^12,¹³ (Table 64.1).

6 Monotherapy with a single agent effective against the expected organisms aims to decrease the risk of drug antagonism, reaction or toxicity.^14,¹⁵ Monotherapy often costs less than multiple antibiotic usage.

7 The clinical response to treatment already given should always be considered, when bacteriological results suggest a change in antibiotics.

8 A standard 2-week course of antibiotics is unnecessary and probably harmful. There is a move to use shorter courses for pneumonia ^16,¹⁷ (see Table 64.1).

9 In consultation with infectious disease specialists, additional tests such as antibiotic minimum inhibitory concentration (MIC), antibiotic assay, serum bactericidal activity and synergy tests of antibiotic combinations may be useful in serious infections (e.g. endocarditis and infections in immunocompromised patients). Sensitivity tests should be interpreted carefully. In vitro sensitivity does not equate with clinical effectiveness; in vitro resistance generally predicts clinical ineffectiveness.

10 Consultations with the laboratory staff and infectious diseases/clinical microbiology specialists are always useful and should be mandatory in serious infections (e.g. meningococcal sepsis, meticillin-resistant staphylococci and multiresistant Enterobacteriaceae).

11 The pharmacokinetics and pharmacodynamics (e.g. penetration into relevant tissues) as well as the spectrum of activity of the antibiotic must be considered. Antibiotic pharmacokinetic principles should determine the dosage and frequency of antibiotic regimens (see below).

12 Adequate drug doses should be given. The intravenous route is preferable in critically ill patients, but other routes should be considered when appropriate (e.g. rectal metronidazole ¹⁸).

13 Serum levels of potentially toxic antibiotics should be monitored, especially if hepatic or renal dysfunction is present.

14 Prophylactic use of antibiotics should be limited to certain situations, should cover organisms that can potentially cause infections in that specific group of patients (e.g. organisms causing skin and soft-tissue infections differ from those implicated in intra-abdominal infections) and should be given at the appropriate timing (see below).

15 General signs of infection are signs of systemic inflammation. Although bacterial infection is likely, non-bacterial infection and non-infective causes should also be considered. High procalcitonin and C-reactive protein levels may be discriminatory for infection,¹⁹ as are interleukin-6 levels, although these are difficult to measure.

16 Antibiotic guidelines are only one aspect of infection control.²⁰ Hand-washing and hand hygiene in general are vital and fundamental aspects of infection control.²⁰ Identification and elimination of reservoirs of infection, blocking transmission of infection, barrier nursing, interrupting progression from colonisation to infection and eliminating risk factors such as invasive devices are also important.

Table 64.1 New paradigm of treatment for nosocomial sepsis

Old	New
Start with penicillin	Get it right first time (broad-spectrum)
Cost-efficient low dose	Hit hard up front
Low doses = fewer side-effects	Low dose → resistance
Long courses ≥ 2 weeks	Seldom longer than 7 days

COMMON ERRORS WHEN USING ANTIBIOTICS

1 Administration of antibiotics before microbiological specimens are obtained.

2 Inadequate quantity and poor quality of blood cultures.

3 Extended use of antibiotics after eradication of infection (e.g. a 2-week course for ventilation associated pneumonia ¹⁶)

4 Antibiotic ‘surfing’ (i.e. switch from one combination to another) when patient is not improving without delving into the cause of persistent inflammatory response.

5 Inadequate and delayed therapy and/or incorrect dosing of antibiotics.

6 Failure to predict adequately ‘resident’ microbial flora and therefore inability to choose correctly empiric antibiotics for nosocomial infections, i.e. no adequate surveillance data.

7 Failure to recognise toxic effects of antibiotics, particularly when polypharmacy is used.

8 Use of combination therapy, irrespective of infection.

SPECIFIC ISSUES

PHARMACOKINETIC PRINCIPLES

The goal of antimicrobial prescription is to achieve effective active drug concentrations (a combination of dose and duration) at the site of infection whilst avoiding, or at least minimising, toxicity.

The various antibiotic classes have different ‘kill characteristics’ and therefore should be dosed differently ²¹ (Table 64.2).

Table 64.2 Pharmacodynamic properties of selected antibiotics

β-LACTAMS (ALL PENICILLINS AND CEPHALOSPORINS, MONOBACTAMS)^22,²³

1 Studies of β-lactam antibiotics on Gram-negative bacilli show a bactericidal activity that is relatively slow, time-dependent and maximal at relatively low concentrations. Bacterial killing is almost entirely related to the time that levels in tissue and plasma exceed a certain threshold. The maximum time plasma β-lactams levels should be allowed to fall below MIC is 40% of the dosing interval.

2 β-lactam antibiotics lack a significant postantibiotic effect (PAE), particularly against Gram-negative organisms, and it is not necessary to achieve very high peak plasma concentrations. PAE is the continued suppression of bacterial growth despite zero serum concentration of antibiotic. It is suggested that concentrations of any β-lactam should be maintained at about 4–5 times MIC for long periods, as maximum killing of bacteria in vitro occurs at this level. If antibiotic concentrations fall below this threshold in the in vitro models, bacterial growth is immediately resumed.

3 Thus, it is important for the efficacy of β-lactams that the dosing regime maintains adequate plasma levels for as long as possible during the dosing interval. It is not surprising, then, that dosing regimes of β-lactam antibiotics are being re-evaluated to keep plasma levels above certain thresholds of MICs of Gram-negative organisms for longer periods of the dosing interval.^24,²⁵

CARBAPENEMS

2 Prolonged infusions (over 3 hours) have been utilised to improve time above MIC.^25,²⁶

3 In vitro data suggest low concentrations may predispose to development of resistant organisms.²

AMINOGLYCOSIDES (THE BEST PHARMACOKINETICALLY STUDIED OF THE ANTIBIOTIC CLASSES)^22,²³

1 The above contrasts with the kill characteristic of the aminoglycosides, which is concentration-dependent. Experimentally, a high peak concentration of an aminoglycoside antibiotic provides a better, faster killing effect on standard bacterial inocula.

2 All aminoglycosides exhibit a significant PAE. The duration of this effect is variable, but the higher the previous peak, the longer the PAE. This phenomenon allows drug concentration to fall significantly below MIC of the pathogen without allowing regrowth of bacteria and hence not compromising antibacterial efficacy. The phenomenon of PAE is much more prevalent in aminoglycosides than other antibiotics and more pronounced with Gram-negative bacilli than with other bacteria.

3 These principles allow for single daily doses of aminoglycosides (also termed extended interval dosing). Combining various meta-analyses involving thousands of patients, once-daily administration was found to be more efficacious, with reduced toxicity, higher peak/MIC ratios, further prolonged PAE and reduced administration costs.

4 With renal dysfunction ²⁷ the dose should be altered according to creatinine clearance. If ≥ 60 ml/min, give 5–7 mg/kg daily; if 59–40 ml/min, the same dose at an interval of 36 hours; 39–20 ml/min, increase the dosing interval to 48 hours.

QUINOLONES ^22,²³

1 In contrast ciprofloxacin has a combination of both the above (concentration-dependent and time-dependent effects) as well as some PAE.

2 Although one suggested ‘target’ parameter for a good clinical bactericidal effect is a high peak, the most validated parameter is the area under the inhibitor curve (AUIC), i.e. AUC/MIC: values >125 translate into better clinical cures.

3 There is general concern about the emergence of resistance related to inappropriately low doses of ciprofloxacin ³ (see Table 64.1).

GLYCOPEPTIDES ^22,²³

1 Vancomycin induces a PAE and a postantibiotic sub-MIC effect. These combined effects suggest that bacterial regrowth will not occur for prolonged periods following a fall in drug concentrations to levels below the MIC.

2 Continuous infusions of vancomycin may have some advantages.

Aminoglycosides and glycopeptides distribute well into fluids of the extravascular, extracellular space, and less well into tissues. This has two important implications. Firstly these agents should not be first-line agents for, nor monotherapy for, solid-organ infections (lung, kidney, liver). Secondly, in situations where extravascular fluid shifts are significant, such as in situations of ‘third-space losses’ of abdominal sepsis and in severe burns, the volume of distribution of these drugs is significantly affected. Hence for any serum level required a larger dose may have to be administered. The volume of distribution of the quinolones (very large) suggests that penetration is excellent into most tissues and hence these drugs are good for solid-organ infections. Similarly β-lactams as a group (including carbapenems) all have reasonable tissue penetration.

ANTIBIOTIC PROPHYLAXIS ^4,^6,^7,²⁸

The main indications for prophylaxis are:

1 When surgery involves incision through an area of colonisation or normal commensal flora and a resultant potential infection has morbidity or mortality.

2 When a procedure (e.g. catheterisation, instrumentation, intubation, dental work) potentially produces a bacteraemia in the presence of an immunocompromised patient or when the potential bacteraemia occurs in the presence of an abnormal heart valve or a prosthesis.

Basic principles of choice of prophylactic regimen should include:

1 The organisms colonising the area through which the incision is made should be covered (Gram-positives if skin is breached, Gram-negatives and anaerobes if bowel is opened).

3 If prevalence of a resistant organism in a specific area is high (e.g. meticillin-resistant Staphylococcus aureus (MRSA) in burns units), then those organisms should be covered by the prophylactic regimen.

TIMING AND DURATION OF PROPHYLAXIS

1 Optimal blood levels of antibiotic(s) are needed when the occurrence of the potential bacteraemia occurs, i.e. for surgery, optimal timing of the antibiotic should be at, or just prior to, induction of anaesthesia and skin incision.²⁸

2 For prolonged procedures where bacteraemias are still a potential occurrence a second dose of antibiotic(s) may be considered.

3 There is no extra benefit of postoperative antibiotic prophylaxis.

DOSES

Comments on dosing regimens are provided below. Doses suggested below are intravenous, for a 70-kg adult with normal renal function. All these drugs accumulate with renal dysfunction and modified doses should be used accordingly.

1 Doses for β-lactams vary with each different drug but recent emphasis supports lower boluses with more frequent administration, i.e. 4-hourly versus 8-hourly or bd versus daily. Continuous infusions may become the standard of practice.²⁵

2 Aminoglycosides: tobramycin and gentamicin 7 mg/kg as a loading dose on the first day, followed by 5 mg/kg per day. Amikacin 20 mg/kg as the loading dose followed by 15 mg/kg per day. These doses are the same for adults and children, neonates excluded.

3 Quinolones: ciprofloxacin at least 400 mg bd (up to tds).

4 Glycopeptide: vancomycin at least 2 g/day either as continuous infusion or in divided doses (40 mg/kg per day for children).

5 Carbapenems: meropenem or imipenem: 3 g/day in at least three divided doses.

SURVEILLANCE

Some type of simple laboratory-oriented surveillance which primarily collects data and resistance patterns of microbiological isolates is important. There are a few international projects,¹¹ but each unit should have access to its own such data as there is increasing prevalence of resistant organisms in ICUs. This is complicated even further by different units having differing resistance patterns.¹⁰ Empiric antibiotic therapy must take these factors into account. Some form of surveillance that provides units with their own microbiological data, updatable quarterly or biannually, is therefore beneficial in helping choose empiric and prophylactic regimens that are applicable to any specific unit.¹⁰

MULTIRESISTANT ORGANISMS

Although this chapter is on antibiotics, the point must be made that without good, efficient and effective infection control policies in all areas treating critically ill patients, the spread of multiresistant organisms would be rampant and the control thereof useless.²⁰ Part of these policies should involve attention to good hand hygiene and the use of antiseptic soaps and alcohol-based hand rubs.²⁰ Hands are still the most documented and incriminated mode of transmission of infection. In this regard a decrement in nursing numbers has also been incriminated in outbreaks of infections, possibly due to the time it takes to wash adequately between procedures.²⁰

1 Multiresistant streptococci and vancomycin-resistant enterococci (VRE), although not common in all countries, are an increasing worldwide problem, as is community-acquired MRSA.

2 The prevalence of MRSA is wider, with many ICUs having this organism almost endemic. Community-acquired MRSA is becoming a huge problem, largely because of identification and initial treatment.²⁹ It differs from the hospital-associated strain in that it is far less multidrug-resistant, but nevertheless it is oxacillin-resistant.

3 New agents are available for treatment of resistant Gram-positive infections.³⁰

4 Antibiotic resistance is agent-specific. Often resistance is claimed to be against third-generation cephalosporins, but this is largely to ceftazidime (particularly the extended-spectrum beta-lactamases of Klebsiella pneumoniae, Escherichia coli and some Enterobacteriaceae).

5 Worrying Gram-negative organisms are Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumanii spp. and Stenotrophomonas maltophilia (the latter specifically, for which trimethoprim may need to be used). A common feature of these organisms is intrinsic resistance to multiple antibiotics. P. aeruginosa and Acinetobacter complex (also named A. baumanii) have become particular problems. Sulbactam and polymixin B or colistin have been used for these problem organisms.³¹

6 Since the carbapenems there are no new agents for multiresistant Gram-negative organisms,¹ although tigecycline has recently been launched.

MONOTHERAPY VERSUS COMBINATION THERAPY ^14,¹⁵

1 Much of the work in this area was performed before the clinical introduction of the carbapenems, penicillin/betalactamase combinations and fourth-generation cephalosporins. It seems that newer single agents are adequate apart possibly for resistant pseudomonal infections.¹²

2 There is no clear evidence supporting the claim that combination antimicrobial therapy prevents emergence of resistance.¹⁴

3 However combination therapy is often suggested for endocarditis and some pseudomonal infections.¹² When combination therapy is used, preference should be given to the combination therapy of two different classes of antibiotics and then classes that act synergistically. The combination of two β-lactam antibiotics should not be used.

BROAD-SPECTRUM INITIAL COVER WITH DE-ESCALATION ³² (SEE Table 64.1)

In view of the morbidity and mortality of delayed appropriate therapy for nosocomial sepsis,^12,^13,³² particularly pneumonias, patients with risk factors for infection with resistant pathogens should initially receive broad-spectrum antibiotics, possibly even combination therapy,^12,³² and, as soon as the pathogen and the susceptibilities are available, treatment should be simplified to a more targeted one – so-called ‘de-escalation’ therapy.³² In the limited studies to date, de-escalation has led to less antibiotic usage, shorter durations of therapy, fewer episodes of secondary pneumonia and reduced mortality, without increasing the frequency of antibiotic resistance.