Antibiotic prophylaxis in the era of multidrug-resistant bacteria

Abstract

The prophylactic use of antibiotics can only be justified when clinical benefits on relevant patient outcomes, such as morbidity or mortality, cost-effectiveness, and absence of immediate emergence of antibiotic resistance have been unequivocally demonstrated. In some intensive care unit (ICU) patients, antibiotic prophylaxis is used as part of selective digestive tract decontamination (SDD) and selective oropharyngeal decontamination (SOD). Recent trials in ICUs with low levels of antibiotic resistance strongly suggest that both regimens reduce the incidence of ICU-acquired infections and improve patient survival. Naturally, the unique microbial ecology of such settings reduce generalizability of results. Therefore, the routine use of SOD and SDD remains highly controversial, especially in ICUs with higher levels of antibiotic resistance. Moreover, convincing evidence is still missing on several important aspects related to efficacy and safety. Despite numerous trials, effects of SDD and SOD on antibiotic resistance during and after decolonization treatment have still been insufficiently investigated, and existing results are contradicting. Furthermore, the effects of both regimens on the non-culturable part of the intestinal flora remain unknown. Finally, cost-effectiveness has not been thoroughly investigated, and prices of the antimicrobial agents that have been used have increased dramatically in recent years. In this review, important knowledge gaps that so far prevent the widespread use of SDD and SOD will be addressed.

Keywords::

• antibiotic
• intensive care
• prophylaxis
• resistance
• selective digestive tract decontamination
• selective oropharyngeal decontamination

Introduction

Minimizing antibiotic exposure is the cornerstone of preventing emergence of antibiotic resistance, and therefore an important aspect of prudent and rational antibiotic stewardship. Naturally, this does not include liberal prescription of antibiotics to patients without documented bacterial infection or a strong clinical suspicion of such an infection. As a result, prophylactic use of antibiotics should have clear and undisputed benefits, as in surgical prophylaxis. Selective digestive tract decontamination (SDD) and selective oropharyngeal decontamination (SOD) are prophylactic strategies for patients treated in intensive care units (ICUs) (see Table 1). However, both strategies are highly controversial, mainly because of inconsistent (and frequently non-convincing) results from underpowered single-center studies during the last 30 years. Herein, we discuss the missing links in the evidence base for both strategies to be regarded as safe and beneficial for widespread use in this era of rapidly increasing antibiotic resistance.

Table 1. Components of SOD and SDD.

	SDD	SOD	Medication used
Systemic	X		3rd Generation cephalosporin, (i.e. cefotaxim 1000 mg 4 times daily) during the first 4 days of admission
Mouthpaste	X	X	Tobramycin 2%, amphotericin 2%, colistin 2%; 4 times daily
Intestinal	X		Aminoglycoside (i.e. tobramycin 80 mg 4 times daily) Polymyxin (i.e. colsitin or polymyxin E 100 mg 4 times daily) Antifungal antibiotic (i.e. amphotericin B 500 mg 4 times daily)

SDD: Selective digestive tract decontamination; SOD: Selective oropharyngeal decontamination.

The concept of selective decontamination

Already in the 1970s, it was demonstrated in mice that the intestinal anaerobic bacterial flora offers protection against colonization and overgrowth of so-called potentially pathogenic microorganisms (PPMO)–aerobic gram-negative bacteria, yeasts, and Staphylococcus aureus [1]. This concept was labeled colonization resistance, and its disruption, for instance through administration of antibiotics or decreased immunity, allows overgrowth of PPMO. As a matter of fact, antibiotics are frequently prescribed to critically ill patients, and PPMO are frequently causing ICU-acquired infections that are usually preceded by colonization of the oropharynx, stomach, and more distal parts of the digestive tract. Colonization and overgrowth with PPMO in the digestive tract can be prevented by topically applied non-absorbable antibiotics, preferably without disturbing the anaerobic flora (by being “selective”). This concept was labeled SDD. SOD aims to only eradicate PPMOs from the oropharynx, and is used in addition to standard oral care (e.g. tooth brushing).

Effects of SDD and SOD in ICU patients

Due to their underlying illness, ICU patients are very susceptible to acquiring colonization and infection with PPMO. As a consequence, antibiotic use in ICU is high. The prevalence of antibiotic use among ICU patients is around 60%. Of all ICU patients, 40% receive empiric antibiotic therapy. Once initiated, empiric therapy is often continued for over 7 days despite persistent negative microbiology results [2]. This illustrates the low threshold for antibiotic therapy among ICU patients, facilitating selection of resistant bacteria.

Furthermore, as a result of frequent physical contacts between health care workers and patients, many opportunities arise for cross-transmission of bacteria, especially when adherence to hygiene protocols is suboptimal. From a theoretical perspective, it is obvious that effective elimination of carriage with PPMO may benefit individual patients by preventing infections and other patients by preventing cross-transmission.

SDD was used in ICU patients for the first time in 1984 [3]. Since then, many studies have been performed, but due to small sample sizes and – sometimes – poor methodological quality, the true effects of SDD have not been clearly established. For instance, improved survival was only apparent when results were aggregated in meta-analyses [4], effects on development of antibiotic resistance were conflicting [5-7], and there was considerable heterogeneity in decolonization regimens, outcome definitions, and patient populations that were studied, limiting generalizability of results [4,8].

Krueger et al. were the first to report a reduction in ICU mortality (RR, 0.508; 95% CI, 0.295–0.875) in a prospectively defined subgroup of ICU patients with medium range severity of illness (APACHE-II) scores receiving SDD [9]. Two other recent studies strongly suggest that SDD and SOD improve survival of patients treated in ICUs in the Netherlands [10,11]. All three studies were performed in settings with low levels of antibiotic-resistant bacteria. The largest trial – a cluster-randomized cross-over study – was conducted in 13 ICUs in the Netherlands enrolling almost 6000 patients [10]. Day-28 mortality and ICU-acquired bacteremia rates were statistically significantly lower during SDD and SOD when compared with standard care (in which none of both regimens were used) [10]. (Table 2) Similar findings had been obtained in a smaller, single-center randomized controlled trial also conducted in the Netherlands [11]. Moreover, antibiotic resistance rates, both among isolates associated with ICU-acquired bacteremia and ICU-acquired respiratory tract colonization, were lower during SDD and SOD, at least for the period of the study [12].

Table 2. Results from Dutch decolonization studies in ICU patients.

	SDD compared to standard care			SOD compared to standard care^‡
	ARR	NNT	RRR	ARR	NNT	RRR
Mortality
Day 28 mortality	3.5%^‡*	29	13.0%^‡*	2.9%*	34	11.0%*
ICU mortality	3.4%^‡*	29	15.2%^‡*	2.3%*	43	10.4%*
In hospital mortality	2.7%^‡*	37	8.5%^‡*	3.4%*	29	10.8%*
ICU acquired bacteremia
≥ 1 episode	6.4%^§	16	49.0%^§	4.0%	25	30.8%
HRMO	0.6%^§	170	6.0%^§	OR 1.10 [0.59-2.07]^§	ND	ND
Candida species and other yeast	0.7%^§	152	7.0%^§	OR 1.16 [0.61-2.21]^§	ND	ND
Glucose-non-fermenting Gram-negative rods (GNF-GNR)	1.0%^‡	98	55.6%^‡	0.9%	110	50.0%
Enterobacteriaceae	3.6%^‡	28	82%^‡	1.3%	78	30.0%

*Calculated from OR adjusted for APACHE II score, age, intubation status, reason for admission, and gender.

^‡De Smet et al. [9].

^§De Smet et al. [22]: analysis of patients with ICU length of stay > 3 days from [9].

ARR: Absolute risk reduction; HRMO: Highly resistant microorganism; ND: Not determined; NNT: Number needed to treat; OR: Odds ratio [95% confidence interval]; RRR: Relative risk reduction; SC: Standard care; SDD: Selective digestive tract decontamination; SOD: Selective oropharyngeal decontamination.

Requirements for justification of SDD and SOD

Based on these findings, either SDD or SOD is now standard of care in most ICUs in the Netherlands. However, these decolonization strategies remain highly controversial in other settings with higher levels of antibiotic resistance. First of all, benefits of either of both regimens on relevant patient outcome parameters remain to be determined in such settings, and if demonstrated, ecological safety should be demonstrated as well.

SDD has been associated with overgrowth and higher prevalence rates of gram-positive bacteria [13-15], methicillin-resistant Staphylococcus aureus (MRSA) colonization [5,16], and gram-negative bacteria resistant to specific antibiotics such as cephalosporins [17] and aminoglycosides [7]. On the contrary, SDD has been successfully used as an adjunct in controlling an outbreak with multi-resistant Enterobacteriaceae [18] and carbapenem-resistant Klebsiella pneumoniae [19]. Furthermore, SDD supplemented with vancomycin was associated with lower MRSA colonization [20-22] and lower respiratory tract infection rates [22] in MRSA endemic settings.

Prolonged use of SDD was not associated with increasing resistance rates in settings in Germany, France, and Spain [13,14,23]. The effect of stopping SDD or SOD on antibiotic resistance in individual patients currently remains unknown (Table 3).

Table 3. Requirements for prophylactic use of antibiotics in decolonization strategies.

	Antibiotic resistance level*
Requirement	Low	Moderate-high	Remark
1. Effectiveness demonstrated for relevant outcomes in individual studies	Yes	No	Convincing evidence from three studies in settings with low levels of resistance [9-11]. Aggregated data from other settings do suggest an effect on mortality, without evidence from individual trials [26]
2. Short-term ecological safety demonstrated	Yes	No	No evidence of increased resistance in the short term in settings with low levels of resistance [9-11]. Increased prevalence of colonization with gram positives and MRSA and resistance to cephalosporins during SDD in settings with moderate-high levels of antibiotic resistance [5,13-16]
3. Long-term ecological safety demonstrated	No	No	No increase in resistance rates during SDD in two studies with follow-up for up to 5 years [13,23]. Increased point-prevalence of resistance to cephalosporins after discontinuation of SDD in one multi-center study [17], which was not observed in another study [6]
4. Ecological safety after ICU discharge demonstrated	No	No	No data available (study ongoing^‡)
5. Ecological safety in non-culturable flora demonstrated	No	No	No data available (study ongoing^‡)
6. Cost-effectiveness demonstrated	No	No	No data available (study ongoing^‡)

*Low is defined as < 5% of bacteremia episodes caused by MRSA for S. aureus, VRE for enterococci and extended spectrum beta lactamase or carbapenamase producing strains for Enterobacteriaceae [9-11]). Moderate-high: settings not meeting the criteria for low levels of resistance.

^‡M. Bonten, personal communication.

MRSA: Methicillin-resistant Staphylococcus aureus; SDD: Selective digestive tract decontamination.

Except for antibiotic resistance, there are other relevant aspects of SDD and SOD that have not yet been evaluated. Cost effectiveness has never been thoroughly investigated (i.e. an analysis of costs versus life years gained), although reductions in ICU-related costs during SDD have been reported [16,24]. At the time that the before-mentioned Dutch studies were performed, the daily costs of antibiotics used in SDD and SOD were $12 and $1, respectively [10]. Mainly because of changes in the pricing of amphotericin B, daily costs of commercially available SDD and SOD are currently $235 and $25, respectively (B. Wittekamp, personal communication).

Antiseptics could be a cheaper alternative to antibiotics. However, although antiseptics, such as chlorhexidine mouth washing, have been associated with reduction in respiratory tract infections in ICU patients [25], their effects on more objective outcomes needs to be further established. It remains unknown whether resistance to chlorhexidine will occur during widespread prophylactic use.

Furthermore, a post hoc analysis of the multicenter SDD–SOD study in the Netherlands suggested an ICU-wide ecological rebound effect with ceftazidim-resistant bacteria after discontinuation of SDD [17]. Such effects have not been studied systematically, but will be critical for the acceptance of these strategies. Finally, monitoring of antibiotic resistance has always been based on routinely performed microbiological culture techniques. However, these techniques can only detect the “culturable” part of the intestinal flora, which may represent less than 5% of the total bacterial flora. It remains unknown whether these strategies stimulate selection of resistance genes in the non-culturable part of the flora.

Conclusion

Despite reported benefits of SOD and SDD on patient survival and prevention of ICU-acquired infections, lacking evidence on critical aspects of decolonization strategies underscores the need for additional research on specific outcomes, especially in settings with higher antibiotic resistance rates than Dutch ICUs. The currently available evidence from Dutch studies that SDD and SOD reduce day-28 mortality, ICU-acquired bacteremia rates, and ICU-acquired colonization with antibiotic-resistant bacteria, holds promise for the future, but does not implicate that the same effects will be obtained in other settings.

Expert opinion

Despite intense disputes on the benefits and threats of SDD and SOD during the last two decades, the discussion on the use of SDD and SOD has not yet been closed. A superficial appraisal of the SDD and SOD literature from the last decades might give the impression that these strategies have been considered as religions, with intensivists and clinical microbiologists divided into believers and non-believers. A more critical appraisal will identify important knowledge gaps regarding effectiveness and ecological safety of SDD and SOD. In our opinion, these knowledge gaps currently prevent the widespread use of SDD and SOD outside settings with bacterial resistance ecology comparable to those in Dutch ICUs. Moreover, prophylactic use of antibiotics as part of SDD and SOD strategies does not justify a more liberal interpretation of antibiotic prescription or standard hygiene policies, as these remain critically important in the prevention of development and spread of (resistant) microorganisms.

In the Netherlands, these regimens are now widely used in ICUs, and studies are ongoing to determine antibiotic resistance development after discontinuation of prophylaxis, both in ICUs and in individual patients, and effects of SDD and SOD on resistance development in the non-culturable part of the intestinal flora. Furthermore, recent cost developments of amphotericin B necessitate a formal cost-effectiveness analysis.

The beneficial effects as reported in Dutch studies have not been obtained in settings with different bacterial ecology, that is, with higher levels of antibiotic-resistant bacteria. There are several initiatives to investigate SDD and SOD in such settings. These include a European-wide multicenter trial in which the effects of SDD and SOD but also of oropharyngeal decontamination with chlorhexidine that will be tested in ICU patients (B. Wittekamp, personal communication). Results of these and other studies should help to complete our understanding of all relevant aspects of decolonization strategies using (prophylactic) antibiotics in ICU patients.

Declaration of interest

The authors state no conflict of interest and have received no payment in preparation of this manuscript.