De-escalation as a Critical Strategy in Acute Care Prof Bojana Beović, MD, PhD University Medical Centre Ljubljana Faculty of Medicine, University of Ljubljana Ljubljana, Slovenia
Disclosures Speaker for Alkaloid, Astellas, AstraZeneca, Altamedics, Lek (Sandoz), MSD, Pfizer Consultant: Lenis, Lek (Sandoz) Research grants: MSD, Pfizer
De-escalation: Definition De-escalation from latin scala, ladder or scalae pl., stairs De-escalation is replacing the empirical antibiotic treatment with a narrower spectrum antibiotic discontinuing redundant antibiotic therapy (switching to oral therapy). www.google.si/search?biw=1067&bih=490&tbm=isch&sa=1&ei=3kldwvknmph9kwx0uihg&q=ladder+falling+from&oq=ladder+falling+from&gs_l=psyab.3...72452.75344.0.75626.13.12.0.1.1.0.152.1434.0j12.12.0...0...1c.1.64.psy-ab..0.12.1344...0j0i67k1j0i8i30k1j0i30k1.0.h68po2ruspi#imgrc=8rr8r1l_ncvkqm:, www.perseus.tufts.edu/hopper/text?doc=perseus:text:1999.04.0059:entry=scala:
The Theoretical Concept of De-escalation Timely broadspectrum antibiotic treatment Successful control of infection (morbidity, mortality) pathogen isolation (exclusion of a pathogen) Narrower spectrum adapted to the isolate and its susceptibility Limited antibiotic selection pressure (lower antibiotic cost)
The Theoretical Concept of De-escalation Timely broadspectrum antibiotic treatment Successful control of infection (morbidity, mortality) pathogen isolation (exclusion of a pathogen) Narrower spectrum adapted to the isolate and its susceptibility Individual patient benefit Limited antibiotic selection pressure (Lower antibiotic cost) Societal benefit: safeguarding the activity of empirical treatment
A: good evidence to support recommendation II: evidence from 1 well-designed clinical trial, without randomization; from cohort or case-controlled analytic studies; from multiple time-series; or from dramatic results from uncontrolled experiments. Clin Infect Dis 2007; 44: 159-77.
There is no separate recommendation for de-escalation, but: Prospective audit and feedback can address de-escalation of antibiotics and duration of therapy. Rapid diagnostics coupled with ASP involvement was also associated with more rapid appropriate de-escalation. PCT as a tool for de-escalation. Barlam T, et al. Clin Infect Dis 2016;62(10):e51 e77.
Evidence based recommendation classified as relevant: 1. Provide rationale for antibiotic start 2. Perform appropriate microbiological sampling 3. Prescribe empirical therapy according to guidelines (day 1) 4. Review diagnosis 5. Evaluate de-escalation based on microbiological results (day 2 to 5) 6. Consider discontinuation of antibiotic treatment (day 3 to 5). Mutters NT, et al. Int J Antimicrob Agents 2018;51:65-70.
Skills of an ID Physician as the Antimicrobial Stewardship Programme Leader Ostrowsky B, et al. Clin Infect Dis 2018; 66:995-1003.
De-escalation: a key pillar to antibiotic stewardship YES!! JF Timsit, F De-escalation of antibiotic therapy Much ado about nothing J Schouten NL
De-constructing De-escalation Not enough evidence: - the damage to microbiota: when and how long - is sequential treatment beneficial or harmful - the role of combination treatment by Huttner B, Pulcini C, and Schouten J. - the role of dosing and duration for resistance selection At present de-escalation is more a woodoo than science! Huttner B, et al. Clin Microbiol Infect 2016; 22:958-9.
Evidence behind IDSA Guidelines (January 1994 to May 2010) A recommendations (good evidence for support): 23% were level I ( 1 randomized controlled trial), 37% were level III (based on expert opinion only) Lee DH, Vielemeyer O. Arch Intern Med. 2011;171(1):18-22. www.google.si/search?q=voodoo+doctor+cartoon&tbm=isch&tbs=rimg:czjxrpnwi1jxijhez-qrocaqvdev9hjkjgaceuiztywv_19onknlxjxfv_19o68xpibhjlz1r8pd4epc0_1-0p6ipey0yoscutp6qs4ibbuesnquvcyheeykhijn6_10emqobpwrfu9haupvxuoqegks6lnplc_1_12hgmav6pr8-psyoscq0qevend9x_1eqceb53c0dfhkhij2jrxemjscksrk9why5ini4mqeglnvhyl3gslzrgst_1yq6lyduiosct_17snqi97lteblv6u7y_1cu- &tbo=u&sa=x&ved=2ahukewjrkuqqordaahubi1akhzu5c5mq9c96bagaebs&biw=1067&bih=490&dpr=1.5#imgrc=g7ru08kmjkyzkm:
Re-constructing De-escalation (a practice oriented approach) Efficacy of de-escalation: Decrease of antimicrobial resistance Less infections with resistant pathogens Less adverse events related to antibiotic therapy incl. C. difficile infections Lower cost of treatment Safety of de-escalation: Morbidity, length of hospital stay Mortality
Re-constructing De-escalation (a practice oriented approach) Efficacy of de-escalation: Decrease of antimicrobial resistance Less infections with resistant pathogens Less adverse events related to antibiotic therapy incl. C. difficile infections Lower cost of treatment Safety of de-escalation: Morbidity, length of hospital stay Mortality
Efficacy of De-escalation: Antibiotic Resistance (Systematic Reviews) The author, year of publication Gutiérrez-Pizarraya, 2017 Guo, 2016 Paul, 2016 Ohji, 2016 Tabah, 2016 Inclusion criteria Randomized or prospective, adult immunocompetent patients with sepsis/septic shock in ICU Adult patients with sepsis/septic shock Adult patients with microbiologically documented infection Any comparative studies that assessed the deescalation therapy ICU, studies assessing the effects or determinants of de-escalation Definition of deescalation -withdrawal of at least one antibiotic -switch to narrower spectrum -both withdrawal of at least one antibiotic -switch to narrower spectrum -withdrawal of an antibiotic in combination treatment -switch to narrower spectrum -withdrawal of at least one antibiotic -switch to narrower spectrum -both variable Number and type of studies included 1 RTC 1 prospective 1 RTC 4 prospective 4 retrospective 3 RTCs 16 observational Assessment of efficacy ND ND 23 studies ND 2 RTC 4 prospective observational 8 retrospective Contradictory results or no effect No effect on antimicrobial resistance found in two studies, Paul M, et al. Clin Microbiol Infect 2016;22:960e967., Ohji G, et al. Int J Infect Dis 2016;49:71 9., Guo Y, et al. Heart Lung 2016;45:454e459. Gutiérrez- Pizarraya A, et al. Expert Rev Clin Pharmacol 2017. DOI: 10.1080/17512433.2017.1293520., Tabah A, et al. Clin Infect Dis 2016;62:1009 17.
The Efficacy of De-escalation: Antibiotic Resistance (recent studies) Author/year of publication Type of study Type of patients Results Lee/2017 retrospective Community-onset bacteremia More blood-stream infections with MDR Gramnegative bacteria in non-de-escalated group, NS. Weiss/2016 retrospective VAP Less ESBL acquisition in the de-escalation group, NS. De Bus/2016 Rattanaumpawan/2017 prospective observational ICU Higher cumulative incidence of bacteria resistant to initial beta-lactam antibiotic or MDR in deescalated group, NS, no difference when only documented infections or only therapies > 5 days were compared. RCT ESBL infections No difference in stool colonisation with MDR bacteria. VAP, ventilator associated pneumonia, RCT, randomized controlled trial, ICU, intensive care unit, HAP, hospital acquired pneumonia, MDR, multidrug resistant Lee CC, et al. Int J Antimicrob Agents 2017;50:371 6.,, Kim JW, et al. Critical Care 2012, 16:R28., Weiss E, et al. Intensive Care Med 2016;42:2098 2100., Rattanaumpawan P, et al. BMC Infect Dis 2017;17:183.
Author/year of publ. The Efficacy of De-escalation: Clostridium difficile Infections Type of study Type of patients Clostridium difficile Infections Snyder/2017 retrospective Patients after HSCT no difference Viasus/2017 prospective observational Community-acquired pneumonia no difference Leon/2014 RCT Severe sepsis in ICU no difference Bohan/2016 retrospective Health-care associated pneumonia de-escalation not associated with C. difficile infection in 30 days. Snyder M, et al. OFID 2017. DOI: 10.1093/ofid/ofx226., Viasus D, et al. J Antimicrob Chemother 2017;72:547 53., Leon M, et al. Intensive Care Med 2014;40:1399 408., Bohan JG, et al. OFID 2016. doi: 10.1093/ofid/ofw244.
The Efficacy/Safety of De-escalation: Superinfections No difference in superinfection rate in de-escalated and not de-escalated arm in a secondary analysis of a large VAP study. No difference in recurrent pneumonia rate in VAP patients in a retrospective study. Non-significantly lower superinfection rate in de-escalation arm in a randomized controlled study investigating de-escalation from 2nd generation carbapenems to ertapenem in ESBL infections. Significantly higher superinfection rate in de-escalation arm in a randomized controlled study in severe sepsis in ICU (p=0.03). In febrile neutropenia patients, there were significantly less confirmed infections in deescalation arm in a retrospective study (p=0.01). Non-significantly higher hospital infection rate in de-escalation arm in patients with ICU stay > 5 days in a prospective observational study in patients with severe sepsis/septic shock. More secondary blood-stream infections with resistant micro-organisms in not de-escalated arm in a retrospective study on community acquired bacteremia, no statistical analysis. Joffe AR, et al. J Crit Care 2008:23:82 90., Eachempati SR, et al. J Trauma. 2009;66:1343 1348., Leon M, et al. Intensive Care Med 2014;40:1399 408.,., Rattanaumpawan P, et al. BMC Infect Dis 2017;17:183, Lee CC, et al. Int J Antimicrob Agents 2017;50:371 6., Kroll AL, et al. J Oncol Pharm Practice 2016;22:696 701., Garnacho Montero J, et al. Intensive Care Med 2014;40:32 40
Benefits and Un-intended Harms of De-escalation: a Mathematical Model De-escalation reduces the use of high-value drugs and preserves the effectiveness of empiric therapy, while also selecting for multidrugresistant strains and leaving patients vulnerable to colonization and superinfection. Hughes J, et al.
Safety of De-escalation: Mortality (1) Systematic review of 16 observational and 3 RTCs with microbiologically documented infections, published until September 2015: adjusted allcause mortality Paul M, et al. Clin Microbiol Infect 2016;22:960e967
Safety of De-escalation: Mortality (2) Systematic review and metaanalysis of studies on deescalation in patients with severe sepsis/septic shock Systematic review of studies on de-escalation in ICU Any comparative studies that assessed the deescalation therapy, analysis per infection 9 studies (1 RCT) 13 studies (1 RCT) 23 studies (2 RCT) All-cause mortality lower in the deescalation group, the difference is not significant. De-escalation is related to decreased allcause mortality (OR =0.68, CI 0.52-0.88) De-escalation associated with lower mortality in most infections. Guo Y, et al. Heart Lung 2016;45:454e459, Ohji G, et al. Int J Infect Dis 2016;49:71 9., Tabah A, et al. Clin Infect Dis 2016;62:1009 17.
45 40 35 30 25 20 15 10 5 0 De-escalation Associated Mortality: Randomized Controlled Studies * severe sepsis Mortality % * p=0.04 p=0.03 HAP, hospital-acquired pneumonia, ICU, intensive care unit, CAP, community acquired pneumonia, ESBL, extendedspectrum beta-lactamases, COPD, chronic obstructive lung diseases HAP in ICU CAP ESBL infections Severe pneumonia in COPD de-escalation no de-escalation *non-significant Leon M, et al. Intensive Care Med 2014;40:1399 408., Rattanaumpawan P, et al. BMC Infect Dis 2017;17:183., Falguera M, et al. Thorax 2010;65:101e106., Xiao B, et al. Exper Ther Med 2017;13:1485-89., Kim JW, et al. Critical Care 2012, 16:R28.
De-escalation Associated Mortality: Conclusions Lower unadjusted mortality in observational studies may be caused by the fact that de-escalation is performed in less severely ill patients. Adjustments in statistical analyses compensated for the known differences between the de-escalated and not de-escalated groups, but there might be other differences that stimulate physicians in observational studies to de-escalate. RCT are very heterogeneous, not controlled for confounding factors, in two studies the empirical regimen was different, small number of patients. Overall, at present, de-escalation does not seem to be associated with higher mortality, but caution is warranted!
Apparently Poor/Unclear Performance of De-escalations: the Reasons Behind Pathogen? Antibiotic? Infection? Setting? Prescriber?
The role of the Pathogen A 70-years old patient with relaps of rectal carcinoma and a fistula to urinary bladder became febrile, he was put on piperacillin/tazobactam, the next day Streptococcus intermedius grew from urine culture: would you de-escalate? A 40-years old patient with sudden onset of lobar pneumonia, co-amoxiclav started, many Gram-positive diplococci and leucocytes in sputum: would you de-escalate? De-escalation was performed in 61% of patients with VAP documented by BAL and 21% of patents with VAP documented by tracheal aspirate. Giantsou E, et al. Intensive Care Med 2007; 33:1533 40.
The role of the Pathogen De-escalation can have poorer outcome if the isolates are not the causative agents of infections or not the only causative agents. Situations related to lower de-escalation rate in observational studies: Infections with MDR bacteria High risk of un-diagnosed pathogen (abdominal infections) Tabah A, et al. Clin Infect Dis 2016;62:1009 17., Paul M, et al. Clin Microbiol Infect 2016;22:960e967
The Role of Antibiotics No standard protocol for de-escalation in the studies: decreasing the number of antibiotics, change to narrow spectrum, discontinuation. The time to de-escalation in the studies varies from 2 to 7 days. Various definitions of narrow and broad spectrum antibiotics. Tabah A, et al. Clin Infect Dis 2016;62:1009 17., Xiao B, et al. Exper Ther Med 2017;13:1485-89.
Ranking of Antibiotics According to Their Antimicrobial Spectrum: a Tool to Guide/Assess the De-escalation Madaras-Kelly K, et al.bmc Infect Dis (2015) 15:197. Method: National VA susceptibility data and the Delphi process. Criterium for scoring: susceptibility. Scoring of beta-lactam antibiotics for Gram-negative bacteria (simplified): Score E. coli K. pneumoniae P.aeruginosa 4 p/t, carba, all cephalo 3 aminopeni/ blinh 2 aminopeni 1 p/t, carba, all ceph, aminopeni/ bl-inh p/t, anti-ps carba, anti Ps cephalo Weiss, E, et al. Clin Microbiol Infect 2015; 21: 649.e1 649.e10 Method: Delphi process. Criteria for ranking: susceptibility and ecological impact. Ranking of beta-lactam antibiotics for Gram-negative bacteria (simplified): Rank Antibiotic 1 amino-peni 2 amino-peni/ bl-inh 3 3 th cephalo 4 p/t, 4 th cephalo, anti-ps cephalo 5 ertapenem 6 Imipenem, meropenem, doripenem p/t, piperacillin/tazobactam, amino-peni, aminopenicillins, bl-inh, beta-lactamase inhibitor, carba, carbapenem, cephalo, cephalosporin, anti-ps, antipseudomonal
Step 1 Non-anti-pseudomonal penicillins Second generation or third generation non-anti-pseudomonal cephalosporins Fluoroquinolones : levofloxacin (500mg q12h), moxifloxacin (500mg q24h) Trimethoprim/sulfamethoxazole (160mg/800mg q12h) Step 2 Anti-pseudomonal penicillins: piperacillin-tazobactam (16g over 24h) Third generation anti-pseudomonal cephalosporins: ceftazidime (6g over 24h) Fluoroquinolones: ciprofloxacin (400mg q8h) Step 3 Anti-pseudomonal carbapenems: meropenem (3g over 24h) PRE-DEFINED ANTIBIOTIC RANKING Step 4 Anti-pseudomonal carbapenems (meropenem) + other antibiotic with Gram-negative coverage
Other Antibiotic/Bacterium Issues Related to De-escalation Various resistance selection potential of antibiotics. Co-selection. Resistance may develop early during the antibiotic course but is more pronounced in prolonged treatment. Discontinuation of redundant antibiotics is probably the most sensible de-escalation strategy (the patient is not exposed to another antibiotic). Samer A, et al. Antimicrob Agents Chemother 2005;49: 2543 5., Radberg G, et al. Antimicrob Agents Chemother 1990;34:2142-7., Armand Lefevre L, et al. Antimicrob Agents Chemother 2013; 57:1488 95.
Insufficient PK/PD Target Attainment in De-escalated Antibiotic Regimens in Critically Ill: In silico model based on PK in criticall ill patients and EUCAST MIC distribution 120 100 80 60 40 20 0 The probability of the target attainment EI, extended infusion, I, infusion S. aureus E. coli K. pneumoniae PK/PD targets: 40% time > MIC for carbapenems 50% time > MIC for penicillins 65% time > MIC for cephalosporins Carlier M, et al. Antimicrob Agents Chemother 59:4689 4694.
Example: Other Pharmacokinetic Considerations Penetration of Antibiotics into Pancreatic Tissue: Aminoglycosides: poor penetration Ureidopenicillins, higher cephalosporin generations: moderate penetration, good spectrum Carbapenems, fluoroquinolones, metronidazole: good penetration Howard TJ, Temple MB. Am Coll Surg 2002; 195:759 767.
Less Effective Antibiotics? Vancomycin less effective than anti-staphylococcal penicillins in Staphylococcus aureus bacteremia. Tigecycline less effective than beta-lactam antibiotics in severe infections. Aminoglycosides as monotherapy less effective than beta-lactam with the exception of urinary tract infection. Little information on superiority (or inferiority) of new agents because of non-inferiority design of registration studies. Shen F, et al. Int J Infect Dis 2015;39:25-33., Gonzales C, et al. Clin Infect Dis 1999;29:1171 7., Vidal L, et al. J Antimicrob Chemother 2007;60:247 57.
Infectious Syndromes/Patients The studies on de-escalation included patients with: HAP VAP Sepsis/septic shock Community-acquired bacteremia CAP COPD UTI Intraabdominal infections Pneumococcal bacteremia Infections with ESBL producing bacteria Neutropenic patients Patients with cancer Patients with fungal infections Open questions and controversies in de-escalation related to infections/syndromes: The role of clinical certainty. Less de-escalation in possible multiple concurrent infections (pneumonia or/and UTI in ICU patient or in elderly). Do we need to expand deescalation to out-patients?
The Setting Differences in definition of narrow/broad spectrum reflect the resistance patterns and the perceived ecological impact in a given setting. A change from a carbapenem to 3 rd generation cephalosporin is not de-escalation in high ESBL setting. De-escalation is difficult (and less often) performed) in settings with high resistance rates. Tabah A, et al. Clin Infect Dis 2016;62:1009 17., Huttner B, et al. Clin Microbiol Infect 2016; 22:958-9.
De-escalation: The Role of Prescriber Wide range of de-escalation rate described in observational (ICU: 10 to 60%, outside ICU: 5 to 55%). Different attitudes towards de-escalation in young doctors in training internationally and among specialties. Several studies described the increase in de-escalation rates by the use of various stewardship interventions incl. electronic alerts. Some authors report on escalations based on susceptibility results that should lead to de-escalation. Paul M, et al. Clin Microbiol Infect 2016;22:960e967., Ohji G, et al. Int J Infect Dis 2016;49:71 9., Guo Y, et al. Heart Lung 2016;45:454e459. Gutiérrez-Pizarraya A, et al. Expert Rev Clin Pharmacol 2017. DOI: 10.1080/17512433.2017.1293520., Tabah A, et al. Clin Infect Dis 2016;62:1009 17., Beovic B, et al. ECCMID 2017., Kim M, et al. PLOS ONE DOI:10.1371/journal.pone.0160551., Donaldson AD, et al. J Hosp Infect 2010; 74:304-5.
Practical Tips for Safe (and Effective) De-escalation Consider de-escalation when there is high impact of empirical treatment and high need to preserve its efficacy. Consider de-escalation when second antibiotic therapy with lower resistance selection potential with respect to the setting is available. Make proper clinical diagnosis of infection. Assure relevant microbiological diagnostics. Use rapid diagnostics to accelerate the de-escalation and shorten the exposure to first antibiotic. Second antibiotic therapy should be given in a dose that attains the PK/PD target and the therapeutic concentration in the tissue. If possible: discontinuation of an antibiotic is more effective than de-escalation.
Conclusions The main objective of de-escalation is the control of antimicrobial resistance. According to the current level of evidence deescalation is a complex procedure that requires support of an experienced antimicrobial steward. The resistance rates in many settings world-wide urgently call for lower use of critically important antibiotics and proper de-escalation is one of the methods that may help us to achieve this goal.