Strategies for De-Escalation and Optimized Antibiotic Therapy Guillaume Béraud José Ramón Paño University Hospital of Poitiers

Meet-the-Expert Session

Strategies for de-escalation and optimized antibiotic therapy Guillaume Béraud José Ramón Paño University Hospital of Poitiers. Hospital Clínico Universitario. Zaragoza. Spain France @joserrapa ESCMID [email protected] © by author Outline 1. Antibiotic de-escalation (ADE): flaws & strengths

2. Antibiotic prescriber & ADE: decision making Case-scenarios

3. Antibiotic stewardship & ADE: strategies ESCMID eLibrary © by author Antibiotic de-escalation (ADE): flaws & strengths

ESCMID eLibrary © by author • streamline ESCMID eLibrary• narrow © by author 2007

Supplements to the core active antimicrobial stewardship strategies (based on local practice patterns and resources) A. Education B. Guidelines and clinical pathways C. Antibiotic cycling D. Antimicrobial order forms E. Combination therapy F. Streamlining or de-escalation of therapy G. Dose optimization H. Parenteral to oral conversion ESCMIDDellit TH, et al. Clin Infect eLibrary Dis. 2007 Jan 15;44(2):159–77. © by author F. Streamlining or de-escalation of therapy Target the pathogen… …as a consequence of…

ESCMIDKim JH, Gallis HA. Am eLibrary J Med. 1989 Aug;87(2):201– 6. © by author 2016

Streamlining or de-escalation of antimicrobial therapy is no longer a prominent strategy • The term “de-escalation” appears just 3 times in the entire document • De-escalation mentioned as a possible metric to evaluate ABS programs ESCMIDBarlam TF et al. Clin Infect Dis.eLibrary 2016 Apr 13;62(10):e51–e77. © by author 2016

Why did de-escalation fall- from-grace? ESCMIDBarlam TF et al. Clin Infect DiseLibrary. 2016 Apr 13;62(10):e51–e77. © by author Why did de-escalation fall-from-grace? 1. De-escalation = Hazy concept

• 77 yo male admitted because of acute myocardial infarction. On his hospitalization day #10 (ICU) he develops a septic shock in the setting of ventilator-associated pneumonia. • He is started empirically on Vancomycin + Meropenem + amikacin • On day #13 bronchial aspirate and blood cultures show pure ESCMIDgrowth of wild-type Pseudomonas eLibrary aeruginosa © by author Why did de-escalation fall-from-grace? 1. De-escalation = Hazy concept

#10 #13 • De-escalation or shortening the duration of antimicrobial therapy? • ESCMIDBoth? eLibrary © by author Why did de-escalation fall-from-grace? 1. De-escalation = Hazy concept • 77 yo male admitted because of acute myocardial infarction. On his hospitalization day #10 (Cardiology ward) he develops a febrile UTI (sepsis) • He is started empirically on Piperacillin/tazobactam • On day #13 Urine and blood cultures show pure growth of E. coli: Resistant ampi, amoxi/clav, 1st and 2nd gen ceph, TMP/SMX SusceptibleESCMIDpip/tazo, 3rd and 4th geneLibrary ceph, carbapenem and quinolones © by author Why did de-escalation fall-from-grace? 1. De-escalation = Hazy concept

Piperacillin/tazobactam Ceftriaxone Ciprofloxacin

#10 #13 • Is switching to ceftriaxone, de-escalating? • ESCMIDIs switching to ciprofloxacin, eLibrary de-escalating ? © by author Why did de-escalation fall-from-grace? 1. De-escalation = Hazy concept

• There is NO good definition for “de-escalation”

• Without a sufficiently robust definition, deescalation is a concept difficult to work with ESCMID eLibrary © by author Why did de-escalation fall-from-grace? 2. Has it be proven beneficial* enough?

* Patient outcome * Antimicrobial resistance ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author 2. Has de-escalation be proven beneficial* enough? Patient outcome (mortality)

ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author Why did de-escalation fall-from-grace? 2. Has de-escalation be proven beneficial* enough? Patient outcome (mortality)

ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author 2. Has de-escalation be proven beneficial* enough? Is de-escalation truly associated with decreased mortality? Factors associated with de-escalation (p 0.05-0.002) Positively associated Negatively associated • Initially appropriate empiric antimicrobial therapy • Isolation of a MDR pathogen • Broad-spectrum empiric therapy • Polymicrobial infections • Compliance with national prescribing guidelines • Intra-abdominal infections • Treatment with multiple and “companion” antimicrobials • Positive microbiological cultures • Lower severity of illness scores at Day #0, ADE Day #5 Confounding by indication ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author Why did de-escalation fall-from-grace? 2. Has de-escalation be proven beneficial* enough? Patient outcome (mortality)

• ADE might not be the driver of decreased mortality but, at least, it is safe ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author Why did de-escalation fall-from-grace? 2. Has de-escalation be proven beneficial* enough? 2.1. Patient outcome (mortality) 2.2. Preventing emergence/decreasing resistance “None of the studies were designed to investigate the effect of ADE (or non-ADE) on the acquisition of MDR bacteria” ESCMIDTabah A. Clin Infect Dis .eLibrary 2016 Mar 22;62(8):1009–17. © by author Shall we abandon the de-escalation concept?

ESCMID eLibrary © by author Shall we abandon the de-escalation concept? NO (personal opinion) 1. Absence of evidence ≠ evidence of absence • There is great need to investigate the impact of de- escalation on antimicrobial resistance/patient outcomes SIMPLIFY study

ESCMID eLibrary © by author ESCMID eLibrary © by author Antibiotic resistance and antibiotic use Environment Host Microbiome Bacterium Plasmid Gene ESCMID eLibrary © by author Shall we abandon the de-escalation concept?

ESCMID eLibrary © by author Shall we abandon the de-escalation concept? 2. De-escalation is (still) critical to counteract the culture of (spiraling) empiricism

• The tragedy of empiricism ESCMIDCasadevall A. Clinical Infectious eLibraryDiseases. 2017 Mar 9;64(7):823 –8. © by author Shall we abandon the de-escalation concept? 2. De-escalation is (still) critical to counteract the culture of (spiraling) empiricism RP ResistanceCountry Risk PremiumPremium

Treece

Termany

time ESCMIDCasadevall A. Clinical Infectious eLibraryDiseases. 2017 Mar 9;64(7):823 –8. © by author Shall we abandon the de-escalation concept? 2. De-escalation is (still) critical to counteract the culture of (spiraling) empiricism

• In countries (settings) with a low resistance premium, ADE might be less worthy than in countries (settings) with high resistance premium

ESCMIDCasadevall A. Clinical Infectious eLibraryDiseases. 2017 Mar 9;64(7):823 –8. © by author Is resistance avoidance the only outcome achievable with ADE/Targeted therapy?

Outcome of antimicrobial therapy Priority Clinical (efficacy + adverse effects) • Survival • Restitutio ad integrum +++++ • Symptomatic relief Resistance avoidance • Own patient For the patient: ++/+++++ • Other individuals For the society: +++++ Cost • Patient +/+++++ • Healthcare system +++/++++ Other • Convenience (patient / healthcare staff) +/+++++ ESCMID• Patient satisfaction eLibrary+/+++++ © by author Antibiotic de-escalation (ADE): flaws & strengths

• ADE is an intuitive, poorly defined (definable?) concept

• Uncertain, though plausible, and setting-dependant (RP) ecological benefit…

• … that under certain circumstances seems to be safe (no clinical outcome-associated costs) ESCMID eLibrary © by author Antibiotic prescriber & ADE: decision making

ESCMID eLibrary © by author Antibiotic de-escalation (ADE) is the process of: • changing an ongoing antimicrobial regimen (action) • to another, with which a gain is expected 1. Narrower spectrum -> less potential for emergence of resistance emergence 2. More convenient-> better adherence 3. Better safety profile-> less adverse events 4. Lower cost-> saving 5. (…) Led ESCMID by prescriber physician eLibrary in charge of the patient © by author How decision making works

• Human decisions are NOT (completely) rational as they are subject of cognitive biases, and rules ESCMID eLibrary © by author Decission rules…rules of thumb

1. Loss aversion: doctors (humans) are more prone to avert losses than to achieve gains 2. Inertia: doctors are more prone to remain things unchanged than to change (Status quo bias) 3. (Decision) framing, matters ESCMID eLibrary © by author The four-fold pattern of preferences in decision making (prospect theory)

ESCMID eLibrary © by author The four-fold pattern of preferences in decision making (prospect theory)

High probability Gains Losses (Certainty effect) • Illustrative Prospect • Focal emotion • Risk-associated behaviour Low probability (Possibility effect) ESCMID eLibrary © by author Outcomes of antimicrobial therapy

Outcome Desirability “Assessibility” Clinical • Survival +++++ +++++ • Restitutio ad integrum • Symptomatic relief Resistance avoidance For the patient • Own patient ++/+++++ • Other individuals For the society + +++++

Cost Direct: +++++ • Patient +/+++++ Indirect: +/+++ • Healthcare system +++/++++ Other • Convenience (patient/HC staff) •ESCMIDPatient satisfaction eLibrary+/+++++ +++/+++++ © by author ADE decision making: clear-cut scenarios MSSA CLABSI in a 89 yo male receiving vancomycin (empiric) -> ADE: switch to cefazolin High probability Gains Losses 1.  efficacy 2.  adverse effects

3. Resistance avoidance?

Low probability ESCMIDExpected eLibrarydecision: ADE © by author ADE decision making: clear-cut scenarios Wild type (panS) E. coli BSI (UTI) in a 45 yo woman on ceftriaxone (history of pen allergy) -> ADE: ampicillin High probability Gains Losses 1. Convenience/Comfort

2. Severe adverse event? 1. Resistance avoidance? 3.Litigation?

Low ESCMID probability Expected eLibrary decission: NO ADE © by author Case scenarios (x3)

ESCMID eLibrary © by author Case 1 • A 67 year old woman is hospitalized through ER with 40°C fever, blood pressure is 150/80 • No specific past medical history. • She’s complaining of urge to urinate small amounts of urine, burning sensation when urinating, strong- smelling and cloudy urine, upper back right pain and shivers. No other foci of infection. ESCMID eLibrary © by author Case 1 • Pelvic ultrasonography is normal (no obstacle). • Gram negative bacilli in urine sample and in blood culture. • Ceftriaxone 1g/d IV is started as a probabilistic treatment.

ESCMID eLibrary © by author Case 1 • Would you consider it as an uncomplicated or complicated pyelonephritis? • Would you use another antibiotic, rather than ceftriaxone • Would you add an aminoside? • Would you start of a fluoroquinolone rather than with ceftriaxone? (and which one Ofloxacine or Ciprofloxacine?) ESCMID eLibrary © by author Case 1 • 2 days later, the patient is apyretic, with very mild symptoms • Urine sample: – E. coli fully sensitive: 106 CFU/ml – E. faecalis sensitive to amoxicillin:106 CFU/ml • Blood sample: Same E. coli. ESCMID eLibrary © by author Case 1 • Would you « deescalate »? • How?

ESCMID eLibrary © by author Case 2 • A 67 year old woman is hospitalized with 40°C fever, blood pressure is 150/80, with a systolic murmur (4/6). • She is just recovering from cholecystectomy 7 days before, for which she had a urinary catheter for 3 days. • She’s complaining of urge to urinate small amounts of urine, burning sensation when urinating, strong-smelling and cloudy urine, upper back right pain and shivers. No other foci of infection. ESCMID eLibrary © by author Case 2 • Pelvic ultrasonography is normal (no obstacle). • Gram negative bacilli in urine sample and in blood culture. • Ceftriaxone 1g/d IV is started as a probabilistic treatment.

ESCMID eLibrary © by author Case 2 • Would you consider it as an uncomplicated or complicated pyelonephritis? Or something else? • Would you use another antibiotic, rather than ceftriaxone • Would you add an aminoside? • Would you start with a fluoroquinolone rather than with ceftriaxone? (and which one Ofloxacine or Ciprofloxacine?) ESCMID eLibrary © by author Case 2 • 2 days later, the patient is apyretic, with very mild symptoms • Urine sample: – E. coli fully sensitive: 106 CFU/ml – E. faecalis sensitive to amoxicillin:106 CFU/ml • Blood sample: Same E. coli. ESCMID eLibrary © by author Case 2 • Would you « deescalate »? • How?

ESCMID eLibrary © by author Case 3 • A 67 year old man is hospitalized with 40°C fever, blood pressure is 150/80. • He currently has a urinary catheter (acute urinary retention on a voluminous prostatis) for the last 3 weeks, and a surgery for a prostatectomy is programmed by the urologist within a month. • He’s complaining of burning sensation around the urinary catheter, strong-smelling and cloudy urine, upper back right pain and shivers. No other foci of infection. ESCMID eLibrary © by author Case 3 • Pelvic ultrasonography is normal (no obstacle). • Gram negative bacilli in urine sample and in blood culture. • Ceftriaxone 1g/d IV is started as a probabilistic treatment.

ESCMID eLibrary © by author Case 3 • Would you consider it as an uncomplicated or complicated pyelonephritis? Or something else? • Would you use another antibiotic, rather than ceftriaxone • Would you add an aminoside? • Would you start with a fluoroquinolone rather than with ceftriaxone? (and which one Ofloxacine or Ciprofloxacine?) ESCMID eLibrary © by author Case 3 • 2 days later, the patient is apyretic, with very mild symptoms • Urine sample: – E. coli fully sensitive: 106 CFU/ml – E. faecalis sensitive to amoxicillin:106 CFU/ml • Blood sample: Same E. coli. ESCMID eLibrary © by author Case 3 • Would you « deescalate »? • How?

ESCMID eLibrary © by author Feasibility of ADE: an algorithm Do Microbiology results explain patient’s infection?

YES PARTIALLY/NO (fully) Is antimicrobial Favorable resistance likely? YES course? NO YES NO Remaining Very Feasible ESCMIDVariably feasible eLibraryduration? unfeasible © by author Antibiotic stewardship & ADE: strategies

Tools Efficacy Toxicity Education to assist (Clinical (Clinical prescribing Outcome) Outcome)

Prescriber Patient “Quality” of antibiotic use: ADE Avalilability of experts´ Restrictive Resistance Cost advice interventions (Microbiological (Economical ESCMIDAntimicrobial Stewardship eLibraryOutcome) Outcome) © by author To assist others to make good decisions

1. Contributing to create (and identify) the conditions associated with increased chance of de-escalation (framing)

Sampling 1. Adequate clinical evolution … 2. Inadequate clinical evolution, modification of clinical decisions and therapeutic modification … 3. Revitalized interest: protocol, clinical trials, publications, congresses, reports, …

interest 2 3

24 h 48 h 72 h 3-5 days time ?

Edwards et al. Arch Intern Med 1973; 132:678-82 Courtesy of Dr. Rafael Canton Spencely et al. J Infect 1979; 1:23-26 ESCMID CunneyeLibraryet al. Int J Antimicrob Chemother 2000; 14:13-9 R. Cantón (personal experience) © by author Antibiotic stewardship & ADE: strategies

1. Improving microbiological diagnosis & reporting 1.1. Optimization of microbiological sampling Target: prescribers / nurses Activities: education / check-list / alerts / empowerment 1.2. Rapid diagnostics Target: Microbiology / prescriber Activities: Selection / implementation 1.3. Reporting Target: Microbiology / Prescriber ESCMIDActivities: Selective antibiogrameLibraryreporting © by author Antibiotic stewardship & ADE: strategies

2. Assisting decision making 2.1. Institutionalization of ADE Activities: ADE algorithm / recommendations / antibiotic Time out

2. Assisting decision making 2.1. Institutionalization of ADE Activities: ADE algorithm / recommendations / antibiotic Time out 2.2. Making experts’ advice available (on demand) Activities: Assuring enough number of available experts (France: 3.6 FTE/1000 beds) 2.3. Prospective audit and feed-back ActivitiesESCMID: “Bacteremia Team” eLibrary / MDR or C. diff alerts © by author Conclusions (i): ADE • ADE is an intuitive, poorly defined (definable?) concept

• Uncertain, though plausible, and setting-dependant (RP) ecological benefit…

• … that under certain circumstances seems to be safe (no clinical outcome-associated costs) ESCMID eLibrary © by author Conclusions (ii): prescriber

• ADE is decision-making -> active change with difficult-to-assess (sometimes conflicting) outcomes

• Human decision-making is NOT fully rational

• ASP should assess how priority is ADE in its institution/ward

• Framing ADE (passive supporting): Create (and identify) the conditions associated with sucessful ADE