1. Antimicrobial Stewardship Update
Vasilios Athans, PharmD
Kathleen Degnan, MD

2. Background
Forum for discussing information on new antimicrobials, changes to antimicrobial breakpoints, other stewardship issues
Quarterly update
Will post slides to Stewardship website under ‘Education’ for reference

3. Outline Antimicrobial shortages
New Fluoroquinolone breakpoints and dosing
New Daptomycin breakpoints and dosing
MERINO trial impact
Fosfomycin for non-E.coli urinary tract isolates
Coming soon: rapid blood culture panel for gram positives

4. Antimicrobial Shortages

5. Prevalence & Severity of Drug Shortages
survey sent to pharmacy administrators at 719 acute care facilities
498 (69%) respondents reported managing >50 shortages in the preceding year
664 (92%) respondents reported having <1 month from notification  active shortage
Medication Category
Reporting Shortage (N = 719)
Analgesics/opioids
705 (98%)
Antimicrobials
651 (91%)
Non-opioid anesthetics or sedatives
627 (87%)
Cardiac Medications
610 (85%)
Chemotherapeutics
452 (63%)
Intravenous fluids
375 (52%)
JAMA Internal Medicine 2019; doi: /jamaintermed

6. Mitigation & Allocation Strategies
Mitigation Strategies
Reporting Shortage (N = 719)
Maintain a shortage list
664 (92%)
Substitute with similar drugs
606 (84%)
Stockpiling available supply
585 (81%)
Centrally limiting distribution
572 (80%)
Designating a ‘shortage management’ pharmacist
241 (34%)
Creating a standing shortage committee
259 (36%)
Allocation Strategies
Designating indication-specific distribution
580 (81%)
Buying from alternative suppliers
482 (67%)
Requesting supply from nearby hospitals
411 (57%)
Transferring patient(s) to other hospitals
37 (5%)
JAMA Internal Medicine 2019; doi: /jamaintermed

7. ID Physician Perspective
Adverse Impact Reported
Reporting Shortage (N = 358)
Use of broader-spectrum antimicrobials
268 (75%)
Use of more costly agents
209 (58%)
Use of second-line/less effective therapy
161 (45%)
Use of more toxic antimicrobials
132 (37%)
Delayed treatment initiation
54 (15%)
Prolonged hospitalization
43 (12%)
Routine Antimicrobial Use
“Squeezing the Balloon”
Open Forum Infect Dis 2018;5(4): doi: /ofid/ofy068.

8. Antimicrobial Shortage List
Unavailable
Amphotericin B deoxycholate  liposomal amphotericin when indicated
Clotrimazole troches  fluconazole or monitor pre-emptively
Nystatin oral suspension  fluconazole or monitor pre-emptively
Erythromycin (IV)  Azithromycin (IV)
Critical
Aztreonam frozen pre-mixed bags  converted to vials
Minor/Resolved
Ampicillin/sulbactam  back in supply (guidelines updated)
Azithromycin (IV)
Cefazolin
Cefoxitin
Cidofovir
Metronidazole (IV)
Piperacillin/tazobactam

9. Antimicrobial Stewardship Website

10. Fluoroquinolone Breakpoints

11. New Fluoroquinolone Breakpoints
The Clinical and Laboratory Standards Institute (CLSI) updated breakpoints for Fluoroquinolones used for Enterobacteriaceae and Pseudomonas aeruginosa. Breakpoints lowered for both.
Effective February, 2019

12. New Fluoroquinolone Breakpoints- Rationale
Enterobacteriaceae susceptibility to quinolones has declined over decades
New PK/PD data with noted low target attainment with prior breakpoints.
Discordance between old FQ breakpoints and achievable systemic levels.
Ex- Pt with PsA bacteremia with levofloxacin MIC 2. This would be susceptible by 2018 breakpoint, but based on PK data, unlikely to get systemic levels high enough to achieve target AUC:MIC ratio even with high dose levofloxacin
The Cipro susceptibility among 62,550 U.S. isolates (defined by a MIC at ≤1 µg/mL) dropped from 95.2% in 1998 to 81.1% in 2013.
There was not much PK/PD data available when the initial clinical breakpoints were set at the time of product approval. Initial breakpoints were a compromise between breakpoint organization harmonization and an early understanding of PK-PD principles
New PK/PD data with noted low target attainment with prior breakpoints. (Discordance between old FQ breakpoints and achievable systemic PK- even though isolate is susceptible with MIC of 2 no way pt could achieve an efficacious AUC/MIC ratio based on pharmacokinetics). For example- previously someone with bacteremia with PsA and levofloxacin MIC of 2 which was previously in the susceptible range. But with normal CrCl and a goal AUC/MIC ratio of 72 (based on in vivo murine studies and HAP data)- Goal AUC would be 144. AUC= dose/Clearance. Can calculate pt’s Clearance of levofloxacin which = 8.33 L/hr. So AUC = dose/ = dose/8.33. Dose would have to be ~2200 mg which is higher than max dose of 750mg
Wentzler, E. CLSI AST News Update. Jan, 2019

13. Background
Fluoroquinolone efficacy correlates with AUC:MIC (concentration- dependent killing)
Clinical AUC:MIC target based on two efficacy studies of patients with hospital-acquired pneumonia primarily due to GNRs
Monte Carlo simulations of PK-PD target attainment
The PK-PD index associated with efficacy for quinolones is the Area under the concentration-time curve (AUC) to MIC ratio
The committee determined appropriate target AUC:MIC based on two PK-PD studies on HAP and also using data from mouse models.
Non-clinical AUC:MIC target based on mouse thigh infection models
Monte Carlo simulations of PK-PD target attainment to evaluate candidate susceptibility breakpoints for each drug and organism combination

14. PK-PD Target attainment by MIC value for Levofloxacin 750mg IV daily for Enterobacteriaceae
Monte Carlo simulations – calculated percent probabilities of PK-PD target attainment for Levofloxain and Enterobacteriaceae based on clinical data
On the X-axis are the relative distribution of enterobacteriaecae MICs for Levo. On the Y-axis is the percentage of patients reaching PK-PD Target attainment.
For the clinical target (blue line) based on PK-PD data in humans, there is good target attainment (>90%) for levofloxacin through MIC of 0.5 and then starts to drop. Target attainment drops to <20% at old MIC breakpoint of 2.
Red curves are based on PK-PD in mouse thigh models. Clinical data correlates with a 1 log CFU decline in mouse thigh infection models
USCAST

15. Enterobacteriaceae New breakpoint Old breakpoint New breakpoint
Non-clinical targets are based on neutropenic mice thighs
New breakpoint
Old breakpoint
USCAST

16. PK-PD Target attainment by MIC value for Levofloxacin 750mg IV daily for P.aeruginosa
When the committee was selecting breakpoints for PsA, there was noted to be limited clinical data on patients infected with PsA so they focused on the non-clinical PK-PD target associated with a 1-log CFU reduction from baseline in murine infection models (Red triangle line).
Chose breakpoint MIC of 1 for levofloxacin and PsA even though there was a higher probability of PK-PD target attainment when choose MIC of 0.5 (>90%). Why? Because of distribution of susceptibilities. Modal MIC value is 0.5 (most bacteria fall around this). This would place 8.7% of levoflox MIC values in the intermediate category (MIC 1). At MIC of 1, still have a decent percent probability of PK-PD target attainment with 79.5%-85.4%. Breakpoints chosen to contain the vast majority of wild type MICs. So they elevated susceptible breakpoint to </= 1
USCAST

17. Pseudomonas aeruginosa
New breakpoint
Old breakpoint

18. Summary of Breakpoint Changes
Quinolone Breakpoints for Enterobacteriaceae
Antimicrobial
Susceptible
Intermediate
Resistant
Dosing Regimen
Old
New
Ciprofloxacin ≤1
≤ 0.25 2
0.5 ≥4 ≥1
N/A
400mg IV q12h or 500mg PO q12h
Levofloxacin ≤2
≤ 0.5 4 1 ≥8 ≥2
750mg IV or PO q24h

19. Summary of Breakpoint Changes
Quinolone Breakpoints for P.aeruginosa
Antimicrobial
Susceptible
Intermediate
Resistant
Dosing Regimen
Old
New
Ciprofloxacin ≤1
≤ 0.5 2 1 ≥4 ≥2
N/A
400mg IV q8h
Levofloxacin ≤2 4 ≥8
750mg IV or PO q24h

20. Effect of changes on levofloxacin susceptibility
In terms of effects on susceptibility, the biggest impact is for Pseudomonas. Based on national susceptibility data, these updated breakpoints would cause 9% of isolates to be reclassified from sensitive to intermediate and ~7% to be reclassified from sensitive to resistant
S to I ~2%
S to R 1.5%
S to I ~9%
S to R 6.5%

21. Dosing Considerations
New breakpoints only validated for levofloxacin 750mg dose (IV/PO)
Our guidelines reflect that we should now generally use levofloxacin 750mg
Urinary tract infections:
Consider lower dose for susceptible uncomplicated UTIs (250mg?? In our guidelines vs 500mg??)
Can consider using levofloxacin for urinary isolates reported as intermediate. Probably choose higher dose
We have changed our guidelines to reflect the higher dose.
Due to the high concentration of fluoroquinolones in the urine of pts with normal renal function, pts with cystitis caused by isolates reported as intermediate will likely respond to therapy.

22. Summary
Fluoroquinolone breakpoints lowered for Enterobacteriaceae and P.aeruginosa
Validated with levofloxacin 750mg daily so would use this dose
Exception: UTIs

23. Daptomycin Breakpoints

24. Daptomycin Breakpoint Updates (CLSI)
E. faecium
SDD R ≤4 ≥8
Pre-2019 Breakpoints
SDD based on 8-12 mg/kg/d
Serious infections due to E. faecium
ID consultation recommended
Enterococcus spp. S
SDD R ≤1
2-4 ≥8
S based on 6 mg/kg/d
SDD based on 8-12 mg/kg/d
Serious infections
ID consultation recommended
Rationale for the change includes:
-Differential wild-type MIC distribution between E.faecium and Other Enterococcus spp.
-Reports of clinical failure when treating serious E. faecium infection with standard FDA-labeled doses
-PK/PD data suggesting standard doses do not optimize AUC/MIC exposure vs. E. faecium (i.e. low target attainment)
Other Enterococcus spp. S I R
≤ 2 4 ≥8
S based on 6 mg/kg/d

25. Daptomycin MIC Distribution
Daptomycin susceptibility testing at HUP
E-test if only single blood culture isolates
Broth microdilution if multiple blood cultures or heart valve
Global E. faecium MIC Distribution (N=3651)
HUP VRE Bloodstream Isolates (2/2018 – 2/2019)
Challenges with daptomycin MIC reproducibility. E-test MICs ½ to 1 dilution higher than BMD.
Regardless of method, most E. faecium MICs are at least 2 mg/L
E.faecalis MIC distribution is one-dilution lower and there is no evidence to suggest failures at 6 mg/kg (justification for separate breakpoints)

26. Clinical Exposure Targets
Outcomea
fAUC/MIC target
30-day Survival (monotherapy)
27.4
30-day Survival (combo therapy)
20.0
Microbiologic Response
12.3
Methods:
Data were pooled from observational studies assessing outcomes in daptomycin-treated enterococcal bacteremia. Patients who received an additional antibiotic with activity against enterococci and/or a β-lactam antibiotic at any time during treatment were excluded. Daptomycin exposures were calculated using a published population pharmacokinetic model. The free drug area under the concentration-time curve to MIC ratio (fAUC/MIC) threshold predictive of 30-day survival was identified by classification and regression tree (CART) analysis and confirmed with multivariable logistic regression. Monte Carlo simulations determined the probability of target attainment (PTA) at clinically relevant MICs.
Results:
Of 114 patients that received daptomycin monotherapy, 67 (58.8%) were alive at 30 days. Stratifying the cohort by severity of illness, a fAUC/MIC >27.43 was associated with survival in low-acuity (N=77) patients (68.9 versus 37.5%, p=0.006), which remained significant (p=0.026) after adjusting for infection source and immunosuppression. The PTA for a 6 mg/kg/day dose was % when MIC=4 mg/L (i.e., daptomycin-susceptible) and % when MIC=1 mg/L.
aAdjusted for infectious source and immunosuppression
Clin Infect Dis doi: /cid/ciy749.

27. How Does Standard Dosing Perform?
Target: fAUC/MIC >27.43
Methods:
Data were pooled from observational studies assessing outcomes in daptomycin-treated enterococcal bacteremia. Patients who received an additional antibiotic with activity against enterococci and/or a β-lactam antibiotic at any time during treatment were excluded. Daptomycin exposures were calculated using a published population pharmacokinetic model. The free drug area under the concentration-time curve to MIC ratio (fAUC/MIC) threshold predictive of 30-day survival was identified by classification and regression tree (CART) analysis and confirmed with multivariable logistic regression. Monte Carlo simulations determined the probability of target attainment (PTA) at clinically relevant MICs.
Results:
Of 114 patients that received daptomycin monotherapy, 67 (58.8%) were alive at 30 days. Stratifying the cohort by severity of illness, a fAUC/MIC >27.43 was associated with survival in low-acuity (N=77) patients (68.9 versus 37.5%, p=0.006), which remained significant (p=0.026) after adjusting for infection source and immunosuppression. The PTA for a 6 mg/kg/day dose was % when MIC=4 mg/L (i.e., daptomycin-susceptible) and % when MIC=1 mg/L.
Clin Infect Dis doi: /cid/ciy749.

28. Are Higher Doses More Effective?
(VRE Bloodstream Infection)

29. Are Higher Doses Safe?

30. What about S. aureus? No changes to existing susceptibility breakpoint
Staphylococcus spp. S NS ≤1
>1
No changes to existing dosing recommendations
4 mg/kg/d for complicated skin and skin structure infections
6 mg/kg/d for S. aureus bacteremia
≥8 mg/kg/d may be warranted for endocarditis, osteomyelitis, or persistent bacteremia

31. Summary E. faecium Other Enterococcus spp. and S. aureus
Recommended dose of 8–12 mg/kg/d for SDD isolates
Higher doses appear to be safe with monitoring
Other Enterococcus spp. and S. aureus
No change to pre-existing dosing recommendations

32. MERINO Trial

33. MERINO trial
Non-inferiority trial comparing piperacillin-tazobactam vs meropenem for definitive therapy of ESBL E.coli or Klebsiella blood stream infections
ESBL definition: isolates non-susceptible to ceftriaxone
Primary outcome: all cause mortality at 30d after randomization
N= 391
Hoped to answer the question- can pip-tazo be used as a carbapenem sparing therapy in pts with BSI caused by ESBL
ESBL was defined as having an isolate non-susceptible to ceftriaxone. Though still susceptible to pip-tazo
Harris, PN et al, JAMA, 2018;320(10):

34. MERINO Trial
12.3% of pip-tazo patients vs 3.7% of meropenem patients met the primary outcome
Minimal change in odds ratio when despite adjusting for urinary source of infection and Charlson Comorbidity Index
Conclusion: for ESBL E.coli and Klebsiella BSI, pip-tazo was NOT noninferior, compared to meropenem
Based on this data:
Micro lab will suppress Piperacillin/tazobactam susceptibility result on bloodstream isolates of non-ceftriaxone susceptible E.coli and K.pneumonia strains
ID can call the lab and ask for Pip-tazo results
Susceptibility of non-blood isolates will still be reported what to do? Clinically correlate
Limitations:
No extended infusion pip-tazo
Susceptibilities done with E-tests (known issues with this method)
No correlation found between MIC and outcome
All death cases were unrelated to infection but rather due to malignancies or other comorbidities.
HUP/PPMC Micro lab suppresses the piperacillin/tazobactam susceptibility result on bloodstream isolates of E.coli and K.pneumonia strains that are not susceptible to ceftriaxone.

35. Summary
Pip-tazo susceptibility will not reported for ESBL blood stream infections
Unclear what to do with non-blood isolates

36. Fosfomycin Testing of Urine Isolates

37. Fosfomycin Susceptibility
CLSI susceptibility breakpoints exist only for E.coli and E.faecalis. Testing done by disk diffusion
Micro lab will no longer perform or report Fosfomycin susceptibilities for non-E.coli gram negative organisms
Pseudomonas spp, Acinetobacter spp, B.cepacia complex, S. maltophila, S. saprophyticus and S. capitis are considered to have intrinsic resistance to Fosfomycin (inherent or innate resistance)

38. Fosfomycin Susceptibility
Rationale:
Fosfomycin requires a glucose-6-phosphate (G6P)-dependent transporter to enter the bacterial cell and exert activity.
G6P is added to the agar to standardize testing and mimic the physiologic environment.
Human urine does not contain G6P. New data that the in vitro susceptibility testing does not predict in vivo urinary bactericidal activity.
Eric Wenzler. Abstract: “In vitro susceptibility testing of fosfomycin does not predict in vivo urinary antibacterial activity.” April, ECCMID

39. Fosfomycin Susceptibility
MIC values can change by up to 4 doubling dilutions in absence of G6P
Greater impact on non-E.coli species of the Enterobacteriaceae family which have a higher MIC90 than E.coli and often naturally have additional mechanisms of resistance to fosfomycin (FosA- hydrolyzes fosfomycin)
E.Coli naturally have a very low MIC for fosfomycin

40. Summary
Micro will no longer run Fosfomycin susceptibilities for non-E.coli Enterobacteriaceae urine isolates
Bottom line

41. Rapid Blood Culture Panel for Gram Positives

42. Rapid Blood Culture Panel for Gram Positives
Also looks for resistance genes- mecA, mecC, vanA, vanB

43. Rapid Blood Culture Panel for Gram Positives
Will be run for all first time positive blood cultures with a gram positive on gram stain
Ninety-minute run
Hoping to pair with stewardship efforts
Won’t go live for probably a few more months

44. Questions?