The prevention and management of infections due to multidrug resistant organisms in haematology patients

Abstract

Infections due to resistant and multidrug resistant (MDR) organisms in haematology patients and haematopoietic stem cell transplant recipients are an increasingly complex problem of global concern. We outline the burden of illness and epidemiology of resistant organisms such as gram-negative pathogens, vancomycin-resistant Enterococcus faecium (VRE), and Clostridium difficile in haematology cohorts. Intervention strategies aimed at reducing the impact of these organisms are reviewed: infection prevention programmes, screening and fluoroquinolone prophylaxis. The role of newer therapies (e.g. linezolid, daptomycin and tigecycline) for treatment of resistant and MDR organisms in haematology populations is evaluated, in addition to the mobilization of older agents (e.g. colistin, pristinamycin and fosfomycin) and the potential benefit of combination regimens.

Introduction

Multidrug resistant (MDR) organisms have emerged as significant pathogens in expanding haematology and haematopoietic stem cell transplant (HSCT) populations. Resistance to currently available drugs and a limited array of new pharmaceuticals necessitates novel pharmacological and non-pharmacological solutions. We review infections due to resistant and MDR gram-negative organisms, vancomycin-resistant Enterococcus faecium (VRE) and Clostridium difficile, which contribute significantly to the burden of healthcare-associated infections and poorer outcomes in haematology patients 1,2. Furthermore, we examine the role of infection prevention programmes, screening, restricted antibiotic prophylaxis and antimicrobial stewardship in controlling these infections.

Emerging resistant organisms in haematology patients

Multidrug resistant gram-negative organisms

The emergence of MDR gram-negative pathogens has resulted in adverse outcomes in haematology cohorts 1,2. Gram-negative bacilli contribute up to 71% of bacteraemia isolates in some haematology units, and may be responsible for outbreaks of infection in hospitalized patients 3. Neutropenia and malignancy are independent risk factors for resistant Escherichia coli and Klebsiella pneumoniae bacteraemia 4. Haematological malignancy has been identified as a risk factor for bloodstream infections due to extended-spectrum beta-lactamase (ESBL) gram-negative organisms.5,6 In countries with high rates of antibiotic resistance, ESBL or MDR gram-negative organisms contribute up to 13.7% of clinical isolates 5,7. While high level plasmid mediated ampicillin and cephalosporin resistant gram-negative isolates (e.g. TEM, SHV) are widely reported, carbapenem-resistant Enterobacteriaceae (CRE) with a range of underlying resistance mechanisms (e.g. Klebsiella pneumoniae Carbapenemase (KPC)) are emerging in haematology populations, with associated mortality 8.

ESBL colonization has been reported in 3–32% of all HSCT patients 9–13, with the same organism isolated in 6% of bacteraemic patients in a recent German study 13. ESBL bacteraemia in colonized patients has been reported in 2–9% in other studies 10,12, and the relative risk of ESBL bacteraemia in a colonized patient has been estimated to be 4.5 13.

Poor outcomes in MDR gram-negative infections are noted in febrile neutropenia and haematology cohorts 4. MDR Pseudomonas aeruginosa in haematology patients is associated with a mortality rate of 35.8–83.3% 1,14. ESBL E. coli and MDR gram-negative infections are associated with ICU admission and increased mortality in haematology patients 5.

Vancomcyin-resistant enterococcus

Increasing rates of VRE acquisition have been reported in haematology patients, with VRE being responsible for up to 41.1% of all gram-positive bacteraemias 1,15–17, and prevalence in many haematology units being consistent with endemnicity 18. Factors common to haematology populations (neutropenia, central venous access, prolonged length of stay, intensive care unit (ICU) admission, allogenic-HSCT (allo-HSCT), AML diagnosis, antibiotic therapies) are known risk factors for VRE acquisition 18–21. VRE isolation may be a marker of illness severity, especially if detected early in the post-allo-HSCT period 22.

VRE bacteraemia rates in colonized patients range from 0–34% 13,15–17,23,24. In HSCT VRE-colonized patients, risk factors for VRE bacteraemia include vancomycink use following VRE colonization, prolonged duration of neutropenia and immunosuppression 24.

The impact of VRE is significant. Infection is associated with prolonged hospital stay, increased costs, morbidity and mortality 17,20,25. While attributable mortality for VRE bacteraemia is reportedly low (0–8%) 15,16,20,22–24,26, in neutropenic and allo-HSCT patients it has been associated with poorer outcomes, especially if bacteraemia is prolonged or occurs early in the post-transplant period 16,22,27–29.

Clostridium difficile

C. difficile infection (CDI) is increasingly observed in haematology and HSCT patients 30,31, where rates have been estimated to be twice the rates observed in hospitalized non-haematology patients 24. Incidence has been reported as 10–13% in patients with leukemia, 6–27% for HSCT recipients 30–34 and 5–7% for non-leukemic haematology patients 35. Furthermore, up to one third of infections in haematology patients may present as severe disease 36. Hyper-virulent strains (e.g. NAP-1, typically resistant to fluoroquinolone agents) are widespread in Europe and the US, and have been reported in HSCT recipients 30. Treatment outcomes for CDI have been variably reported in haematology patients 34,35, with a recent study suggesting response to metronidazole and vancomycin to be as low as 53.7% and 50%, respectively 36.

Risk factors for CDI have been identified. In patients with acute myeloid leukemia (AML), CDI has been associated with older age, longer duration of antibiotic therapy, ceftazidime use and prolonged neutropenia 31,34,37. CDI in allo-HSCT patients is strongly associated with graft vs. host disease (GVHD). Increased risk of CDI in allo-HSCT recipients is associated with cord blood transplants, total body irradiation and acute grade 2 GVHD 38. CDI has been associated with prior chemotherapy, receipt of broad-spectrum antimicrobial agents and previous VRE colonization in all HSCT recipients 30,31.

Recipients of allo-HSCT are at higher risk for CDI than those who undergo autologous HSCT (auto-HSCT) 30, with an almost a two-fold increase in incidence for allo-HSCT groups compared with auto-HSCT 39,40.

Infection prevention, antimicrobial stewardship and prophylaxis

Non-pharmacological measures are essential for control and management of MDR organisms in haematology and HSCT patients.

Cleaning, isolation and screening

Multimodal strategies including hand-hygiene, environmental cleaning/disinfection, isolation and surveillance form the backbone of effective prevention programmes 41. Hand-hygiene programmes and electronic surveillance systems have been demonstrated in haematology centres to increase compliance and trend toward significant reductions of nosocomial transmission of organisms such as VRE 42,43.

Bleach-based cleaning has resulted in significant reductions in newly-acquired VRE 44, supporting the previous success of intensive infection control measures to prevent VRE acquisition 45. Chlorhexidine-impregnated washcloths have been associated with reduced VRE colonization in ICU patients 46,47. Although one study has not demonstrated a benefit in haematology-oncology patients 48, a recent multi-centre randomized trial (including HSCT patients) demonstrated an overall reduction in MDR organisms, but not bacteraemia with methicillin-resistant Staphylococcus aureus (MRSA)/VRE, with the daily use of chlorhexidine-impregnated washcloths 49. Novel methods for environmental decontamination, such as ultraviolet disinfection 50, hydrogen peroxide vapour 51 and copper alloy surfaces 52 require targeted evaluation in haematology units and cost-benefit analysis before incorporating into standard prevention strategies.

Although the benefits of screening for VRE colonization are widely recognized in haematology populations 41,53, support for routine ESBL screening programmes is limited. One study of neutropenic cancer patients found an ESBL E. coli colonization rate of 31.8%, and a recent multicentre study demonstrated an ESBL faecal carriage rate of 29% 9,12. Notably, there was no link between ESBL carriage and subsequent ESBL bacteraemia or outcome 12. This raises questions regarding the benefits and validity of ESBL screening in this population and the need for studies examining approaches to empirical antibiotic algorithms in ESBL-colonized patients 54.

Antimicrobial stewardship in haematology patients

In large hospitals, host factors and antibiotic exposure, rather than breaches in infection control, have been associated with VRE colonization in haematology patients 55. Antimicrobial stewardship programmes therefore complement infection control strategies. Antimicrobial stewardship programmes are aimed at reducing MDR acquisition with specific recommendations for haematology populations 56,57. Reduction in use of broad-spectrum antimicrobials is vital to reducing CDI and VRE acquisition 21. Given the risks of MDR infection associated with vancomycin therapy 18, vancomycin should be ceased at 48 h in the absence of suspected infection due to gram-positive organisms 58,59.

Antimicrobial stewardship teams with dedicated staff reduce broad-spectrum prescribing, inappropriate vancomycin use and antibiotic resistance in gram-negative organisms 60–63. Improved antimicrobial stewardship, focused on prescribing of high risk antibiotics (third generation cephalosporins, clindamycin, fluoroquinolones) significantly reduces CDI incidence 64. Empiric therapies should be targeted to local epidemiology, which can be facilitated by a stewardship team. Antimicrobial cycling for neutropenic fever is not routinely recommended 57,58,65.

Interventions to modify gastrointestinal flora

To combat primarily gram-negative bacteraemia, selective digestive decontamination (SDD) has been used in some haematology and ICU units 66–68. Aztreonam and colistin have previously been used for this purpose, with no difference in outcomes between the two drugs 69. A multicentre cluster trial of ICU patients given SDD demonstrated decreases in mortality and bacteraemia 70,71. This trial was limited by a short period of follow-up, use in a low multiresistant prevalence setting and restricted broad-spectrum systemic antimicrobial agents 72. A study of oral gentamicin to eliminate carriage of carbapenem-resistant K. pneumoniae reported a success of 66%, but lacked controls and microbiological follow-up for development of resistance beyond 5 months 73. A recent meta-analysis of 35 studies demonstrated no short-term microbiological resistance 74. Nonetheless, concerns of resultant increases in gram-negative resistance mean that SDD is therefore not widely practiced or recommended 75.

Reducing the density of MDR organism gastrointestinal colonization and manipulation of microbiota has been proposed to reduce VRE infection risk. Following antibiotic therapy, the microbiome in allo-HSCT patients demonstrated VRE dominance 76,77. While administration of probiotics has been proposed to eliminate VRE colonization in non-haematology populations 78, risk is recognized in haematology patients, where bloodstream infections have been associated with probiotic therapy 79. The introduction of diverse intestinal microbiota via faecal transplantation, including Barnesiella spp., to heavily colonized mice reduced VRE colonisation 76,77. Clinical study in haematology populations is required.

Faecal microbiota transplantation (FMT), is increasingly used for refractory and recurrent CDI 80,81, and has been shown to be superior to vancomycin therapy in non-haematology patients with recurrent CDI 82. A single case of fulminant CDI in a haematology patient successfully treated with FMT has been reported 83.

Fluoroquinolone prophylaxis

Fluoroquinolone antibiotics have been used for antibacterial prophylaxis in patients with acute leukemia undergoing chemotherapy and allogenic HSCT recipients with profound neutropenia for ≥ 7 days 65,84, and a reduction in febrile neutropenia events, bacteraemia, hospitalization and mortality has been demonstrated 85–87. However, a meta-analysis of only randomized control trials demonstrated no statistically significant mortality benefit in patients with febrile neutropenia 88.

Despite perceived benefits, fluoroquinolone resistance rates exceeding 20% have been reported following uptake of fluoroquinolone prophylaxis in haematology patients 3,86,87,89–93. Significantly increased ciprofloxacin resistance has been reported by the European Centre for Disease Prevention and Control 94. In addition, breakthrough bacteraemias with MRSA, MDR E. coli and P. aeruginosa have been associated with prophylaxis 95. In patients with AML, higher rates of fluoroquinolone-resistant E. coli have been reported, when compared with isolates in patients not administered prophylaxis 96. Notably, the use of fluoroquinolone prophylaxis may lead to subsequent increases in carbapenem prescribing 87,97.

Fluoroquinolone prophylaxis can be targeted and limited by antimicrobial stewardship programmes 57. Cessation of fluoroquinolone prophylaxis in one centre resulted in a decline in fluoroquinolone-resistant Enterobacteriaceae from 85% to 17% (P = 0.0078) 98.

In our experience the cost of antibacterial prophylaxis with regard to evolving resistance is a major concern.99,100 Australian consensus guidelines for management of haematology patients recommended against routine antibacterial prophylaxis for gram-negative bacteraemia, in particular, fluoroquinolones 101.

Pharmacotherapy

To manage MDR infections adequately in haematology patients, new antimicrobial agents, re-consideration of older therapies (Table 1) and combination regimens are necessary.

Table 1.

A description of the pharmacology and clinical utility of colistin, pristinamycin and fosfomycin in haematology patients with mutli-resistant organisms

Antibiotic	Class	Mechanism of action	Formulation Oral bioavailability	Dosing Renal/hepatic adjustment	Clinical utility	Comments regarding treatment of multi-drug resistant (MDR) organisms in haematology patients
Colistin*	Polymyxin	Not completely understood; believed to interact with LPS on gram-negative membrane	i.v., i.m., inhaled	Loading and maintenance dosing calculations required†	MDR gram-negatives, including Pseudomonas spp. and Acinetobacter spp.	•  Renal adjusted dosing safe in haematology and HSCT patients 104,124.
			NA	Renal adjustment required		•  Successful treatment in a matched pair analysis of MDR Pseudomonas aeruginosa; median treatment duration 13 days, resolution of infection in 20/26 (76.9%). Single patient suffered renal failure 105.
						•  Case report/series data of successful colistin single agent and combination therapy for allogenic HSCT patients and haematology patients with MDR Pseudomonas spp 107,149.
Pristinamycin	Streptogramin	Inhibit protein synthesis via binding peptidyl transferase of 50S subunit of 70S ribosomes	Oral Excellent	500 mg−1 to 1 g TDS No adjustment required	Resistant gram-positive cocci (streptococci, staphylococci > enterococci)	•  Limited data specific to haematology cohorts. Majority of data for suppressive therapy for resistant CoNS, MRSA or VRE bone and joint infections 150,151.
						•  In a retrospective review of 46 cases predominately suppression of VRE or staphylococci (58%) joint, bone or wound infections 152.
Fosfomycin	Phosphoenolpyruvate analogue	Bacterial cell wall inhibition by binding to and inactivating enzyme enolpyruvate transferase	Oral or i.v.	•  3 g single dose (simple UTI)	MDR gram-negatives (except Pseudomonas) and gram-positive cocci (S. aureus and Enterococcus spp. except E. faecalis)	• No specific data for use in haematology patients. Usual clinical indication is gram-negative UTIs, soft tissue infections and surgical prophylaxis 153,154.
				•  3 g alternate days (complicated UTI)		•  Primarily for MDR gram-negative UTI, not bacteraemia (E.coli, Citrobacter, Proteus mirabilis including ESBL, however excluding Enterobacter spp., P. vulgaris, Providencia spp., Acinetobacter spp., Pseudomonas spp., Morganella morganii) 153.
			Good‡ (excellent urinary levels)	•  8 g BD i.v. Renal adjustment if CLcr <50 ml min−1		•  In vitro and in vivo synergy with carbapenems, colistin, fluoroquinolones and tigecycline in 30–74% of gram-positive and gram-negative organisms 155,156.
						•  Successful mono or combination therapy for suppression of chronic MDR gram-positive infections (e.g. Staphylococcus aureus, Enterococcus spp.) 153

Adapted from Kucers'. The use of Antibiotics (6th edition) – For ‘Class’, ‘Mechanism of action’, ‘Bioavailability’, ‘Dosing’ sections 124. *Colistin is commercially available in two forms: colistin sulphate (referred to in this publication as colistin) and sodium colistin methanesulphonate (CMS). CMS is ‘less toxic’ when administered intravenously and is present in all parenteral (and most inhaled) formulations. CMS undergoes conversion in vivo to form a mixture of partially sulphonmethylated byproducts and colistin 157. †Colistin dosing dependent upon organism MIC and patient renal function. Loading dose of CBA (colistin based activity)(mg) = Colistin organism minimum inhibitory concentration (mg l⁻¹) × 2 × body weight (kg) (use lower of ideal or actual body weight). Maintenance dose of CBA (mg) = Colistin organism minimum inhibitory concentration (mg l⁻¹) × (1.50 × CL_cr + 30) 158. ‡Bioavailability: 37–42%. Oral formulation has high levels in urine for 1–3 days post single dose (1053–4415 mg ml⁻¹). BD, twice daily; CoNS, Coagulase-negative Staphylococcus; CL_cr, creatinine clearance; ESBL, Extended-spectrum-beta-lactamase; HSCT, haematopoietic stem cell transplantation; i.m., intramuscular; i.v., intravenous; LPS, lipopolysaccharide; MDR, multidrug-resistant; MIC, minimum inhibitory concentration.; MRSA, methicillin-resistant Staphylococcus aureus; NA, not applicable; TDS, three times daily; UTI, urinary tract infection; VRE, vancomycin-resistant enterococcus.

Resistant gram-negative infections

To date, there are no published reports of randomized trials evaluating treatment options for MDR gram-negative infections in haematology patients. Current American and European febrile neutropenia practice guidelines for haematology patients either colonized or with previous MDR Enterobacteriaceae infections suggest empiric colistin and a beta-lactam agent (or one of tigecycline, aminoglycoside, fosfomycin).58,102

Colistin

Colistin (polymixin E) has been clinically available since 1959 and increasingly used as a treatment option for infections caused by MDR gram-negative pathogens 103. However, data on colistin therapy in HSCT and haematology patients are limited. Despite renal toxicity (0–50%), it has been found to be safe in neutropenic patients.104–106 and is recommended for treatment of carbapenem-resistant Enterobacteriaceae and P. aeruginosa 102. Successful treatment of MDR Pseudomonas has been demonstrated in bacteraemic haematology patients, predominately in conjunction with a beta-lactam agent.105,107

Tigecycline

Tigecycline is a first-in-class glycylcycline with activity against MDR gram-negative infections, excluding P. aeruginosa, Proteus/Morganella/Providencia species 108. Clinical utility has been limited by low peak serum concentrations, rapid post treatment resistance and increased mortality and failure rates 109–111. In critically ill patients, tigecycline has been reported to have an overall success of 73%, highest in intra-abdominal infections (82%). In the same study, a 42% failure rate was noted in empirical febrile neutropenia therapy and 30% failure rate in bacteraemia even if the isolate was clinically susceptible 112. In oncology patients (58% haematology, 28% neutropenic) an overall response rate of 64% has been reported 113. A retrospective review of tigecycline as salvage therapy for febrile neutropenia has suggested an overall success rate of 43% 111. Tigecycline is not recommended for empiric febrile neutropenia therapy 58. Meta-analysis and systematic reviews highlight increased mortality and clinical failure. The USA Food and Drug Administration (FDA) recently issued a black box warning 114. Therefore, tigecycline can only be recommended for salvage therapy or as a component of combination regimens for MDR gram-negative infections.109,110

Combination regimens

Successful combination therapies for MDR gram-negative infections have been demonstrated in predominately non-haematology patients and neutropenic gram-negative sepsis. 115–117 Colistin/rifampicin therapy has in vitro and in vivo synergy in small case studies of Acinetobacter baumannii infections 118,119. However, in a randomized trial the addition of rifampicin to colistin offered no benefit 120. Colistin/carbapenem therapy has demonstrated a mortality benefit compared with colistin monotherapy in a retrospective review of MDR gram-negative infections, including haematology patients 121. This was also demonstrated in solid-organ transplant recipients, where colistin/carbapenem therapy for drug-resistant A. baumannii was associated with survival 122. A recent study of patients with carbapenem-resistant (KPC)-producing K. pneumoniae infections, including haematology patients, demonstrated survival benefit with combination therapy (colistin/carbapenem or colistin/tigecycline) without increased toxicity 123. Despite the majority of studies including few haematology patients, the use of combination therapy including colistin/carbapenem should be considered for MDR gram-negative infections.

Resistant gram-positive infections

Daptomycin

Daptomycin is a cyclic lipopeptide bactericidal antibiotic with in vitro activity against S. aureus (MSSA, MRSA, heterogeneous vancomycin-intermediate Staphylococcus aureus; hVISA) and Enterococcus spp. (vancomycin-sensitive enterococci; VSE, VRE).124,125 Infrequent elevations of creatine kinase have been associated with daptomycin, and limitations include the lack of an oral formulation, uncertain dosing (6–10 mg kg⁻¹) and inactivation by pulmonary surfactant. Although not formally evaluated in haematology patients, efficacy for gram-positive infections in neutropenic patients has been reported.126–128 Treatment failure has been reported in central nervous system Staphylococcal infections 129.

Linezolid

Linezolid is a synthetic oxazolidinone antibiotic with excellent oral bioavailability and intrinsic activity against streptococci, staphylococci (MSSA, MRSA, hVISA) and enterococci (E. faecalis, VSE, VRE) 124. Toxicities include cytopenias, peripheral neuropathy and lactic acidosis, with limited toxicity and effect on engraftment observed with use early post allo-HSCT.130,131 A randomized study has demonstrated cure rates comparable with vancomycin (87.3% for linezolid, 82.5% for vancomycin) and fewer drug-related adverse events 132. Appropriate dosing and treatment duration are vital to prevent resistance for infections due to E. faecium, S. epidermis and S. haemolyticus 133.

These two agents are now used increasingly as targeted therapy for VRE bacteraemia in haematology patients. Teicoplanin therapy is also appropriate for vanB-VRE, although vanA-VRE is inherently resistant to this agent, and inducible resistance in vanB-VRE has been reported 134. While safety and tolerability of teicoplanin has been demonstrated in haematology patients with VRE bacteraemia 135,136, the use of empiric teicoplanin in vanB-VRE colonized haematology patients has not been studied, and clinical guidelines do not support this practice.

Daptomycin use for Enterococcus spp. bacteraemia was reported in a recent multicentre retrospective study (dosing: > 6 mg kg⁻¹ day⁻¹, median 8.2 mg kg⁻¹ day⁻¹), with an overall clinical success rate of 89%, together with low rates of creatine kinase elevation (3%) 137.

There are no randomized controlled trials comparing linezolid with daptomycin. One retrospective review of haematology and allo-HSCT patients found no difference in treatment success and outcome in patients treated with these agents 28. Meta-analysis of nine studies demonstrated no difference in microbiological or clinical cure rates, with a trend towards increased survival with linezolid 138.

Clostridium difficile

Fidaxomicin

First line therapy for CDI is oral metronidazole, with vancomycin reserved for severe or refractory cases. Data supporting alternative therapies for CDI are limited in haematology and HSCT patients. Fidaxomicin has been shown to be non-inferior to vancomycin therapy in oncology patients and to be associated with reduced CDI recurrence.139,140 In HSCT recipients with CDI, cure rates following fidaxomicin have been shown to be less than conventional treatments with vancomycin or metronidazole, although studied cases were more severe and likely to have had pre-treatment prior to fidaxomicin 141.

Older agents, pharmacokinetics and pharmacodynamics

In addition to colistin, antimicrobial agents previously curbed because of toxicity profiles have been re-examined as effective therapies for MDR infections. These include fosfomycin and pristinamycin. In particular, the need for effective suppressive therapies for chronic VRE, MRSA and MDR gram-negative infections where cure is impractical has led to the increased use of agents outlined in Table 1.

Pharmacokinetic (PK)/pharmacodynamic (PD) properties of antimicrobial agents may be unique in the haematology population and strategies based on maximizing these principles may form future pathways for treatment of multiresistant organisms. The potential utility of infusional beta-lactam antibiotics for febrile neutropenia has been reported 142,143. Whilst continuous infusions would increase time above mean inhibitory concentrations (MIC) in neutropenic fever when volume of distribution is altered 142,144, prospective studies demonstrating improved cure rates or mortality benefit are still required before widespread implementation. Dosages provided in this review are based upon those used in clinical studies, and these are reflected in published clinical guidelines. However, during sepsis and those critically ill, the PK/PD of drugs such as meropenem are altered 145,146. A recent study demonstrated concentrations of meropenem were below typical gram-negative MIC values and time above MIC shorter in neutropenic sepsis patients vs. non-neutropenic patients 147. Although computer based simulations suggest longer antibiotic infusions give greater time above MIC 146,148, they do not include a significant haematology cohort. To evaluate adequately optimal dosage of antimicrobial agents across the spectrum of patient age, weight, renal and hepatic functions and presence of systemic inflammatory response syndrome, appropriate data must be captured by antimicrobial stewardship programmes and clinical endpoint studies with respect to variation of these variables.

Conclusion

The management of MDR infections in haematology patients is an increasingly complex problem involving pharmacological and non-pharmacological prevention and intervention strategies. The burden of illness arising from MDR gram-negative pathogens, VRE and C. difficile is expanding more rapidly than therapies are becoming available. Increased use of newer antimicrobial agents, such as daptomycin and linezolid, in conjunction with older antibiotics and combination regimens, are effective treatment strategies. Key principles that should underpin multidisciplinary services provided by infectious disease physicians, pharmacists, infection control consultants and haematologists include:

• limiting fluoroquinolone prophylaxis.

• a greater understanding of microbiological complications of SDD before routine use.

• implementing effective infection prevention strategies.

• evaluating the role of ESBL screening and impacts on empirical antibiotic therapy

• antimicrobial stewardship programmes aimed at reducing unnecessary cephalosporin, vancomycin and carbapenem use to limit further emergence of C. difficile, VRE and MDR gram-negatives.

• judicious use of new effective antimicrobials (fidaxomicin, linezolid, daptomycin) against MDR organisms in haematology patients.

• a return to older antimicrobials for MDR organisms (fosfomycin and pristinamycin) and exploration of new combination therapies involving colistin for resistant gram-negatives.

• exploring infusional beta-lactams for neutropenic sepsis to improve time above MIC and patient outcome.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.