Thread by @canuckIDpharm

Our #vancomycin AUC-based therapeutic drug monitoring (TDM) counterpoint is out!

Thrilled to work on this with @wfwrighID @BradSpellberg & Twitterless Andrew Shorr 😀

Check out 🧵for highlights? 👇 1/n

https://twitter.com/ABsteward/status/1373005641531457550

Brief background.
1st 2009 #vancomycin guidelines recommended troughs 15-20 for serious MRSA infections due to #vancomycin ‘MIC creep’ & associated treatment failures.
But not based on evidence for efficacy or safety 😶

academic.oup.com/ajhp/article-a… 2/n

Since then, we learned MRSA #vancomycin ‘MIC creep’ likely due inherent MIC assay variability / limited clonal outbreaks.
academic.oup.com/jac/article/73… 3/n

S. aureus #vancomycin resistance still ++ rare & no ‘MIC creep’ in large surveillance projects.
MIC 90 = 1 for decades
academic.oup.com/ofid/article/6… 4/n

Plus, high #vancomycin MIC link to ↑ treatment failure may NOT be due to #vancomycin under-dosing → Patients with MSSA infections treated with beta-lactams have worse outcomes if isolate has high #vancomycin MIC 🤔

academic.oup.com/jid/article/20… 5/n

So.......high #vancomycin MIC may be marker of patient/pathogen factors related to poor outcomes

Impact of 2009 #vancomycin guideline recommendation to target trough 15-20 →

No improvement in efficacy but ↑↑ AKI😳

aac.asm.org/content/57/2/7…
6/n

💥New #vancomycin guidelines 2020💥
Key recommendation AUC-based TDM for serious MRSA infections, target AUC/MIC 400 – 600.….
or assume MIC = 1, so target AUC 400 – 600 (reasonable based on MIC 90 = 1)

academic.oup.com/ajhp/article/7… 7/n

#Vancomycin AUC TDM ++ resource intensive

Could justify if 2 key assumptions true:

1. Clear relationship between #vancomycin efficacy/toxicity & AUC 400-600

2. Maintaining #vancomycin exposure within AUC 400-600 with AUC-based TDM ↑ efficacy or ↓ toxicity. 8/n

Assumption #1
Lower limit AUC ≥ 400 traced back to retrospective, cohort study by Moise-Broder et al. 2004
link.springer.com/article/10.216… 9/n

N=90 S. aureus pneumonia (37% MRSA)
Successful 'clinical response’ 7 (!) times higher if AUC/MIC ≥350

N=34, faster sterilization of respiratory specimens if AUC/MIC≥400 - Is that an endpoint we care about?

Small study, confounding, combination therapy, mortality?......10/n

Zero (0) studies since have been able to validate AUC/MIC ≥ 400!

Lots have tried!

aac.asm.org/content/57/4/1…

academic.oup.com/cid/article-ab…
link.springer.com/article/10.100…

minervamedica.it/en/journals/mi…

journals.lww.com/drug-monitorin…

minervamedica.it/en/journals/mi…

meridian.allenpress.com/jppt/article-a…
11/n

Other studies used CART to find alternative cut points which ranged from 211 to 667😲

CART = “data torture” 😬to explore exposure-response relationships

1. Infinite # cut points tested to find ‘optimal’ value→ ↑↑↑ type I error
2. Effect size overestimated
12/n

3. Biologically implausible to dichotomize
4. Small change in data can totally change results

See 1st study by Moise-Broder with overlapping patients– cut point was 866 not 400 🤨

academic.oup.com/ajhp/article-a…
13/n

What about relationship between #vancomycin AUC ≥ 600 & AKI?
Again, thresholds range from 350 – 1300 (by CART)
DATA conflicting if trough vs. AUC better AKI predictor

academic.oup.com/cid/article/49…
pubmed.ncbi.nlm.nih.gov/23147106/
aac.asm.org/content/61/5/e… 14/n

Many studies not considered in guidelines show very high correlation (R2 >0.90) or high (R2 0.70 – 0.90) between AUC and trough - See Table S1 in our article. #MythBusting 15/n

📈Bayesian methods seem cool 😎 BUT predictive performance relies heavily on how well the PK model embedded as priors fits the population of interest....
16/n

Check out the variability in the 20+ #vancomycin PK models in this excellent paper.
Which one is best for your patient?

clinicalmicrobiologyandinfection.com/article/S1198-… 17/n

So....no compelling evidence for existence of #vancomycin AUC therapeutic range

Likely highly individual

Using population therapeutic range derived from weak evidence antithesis of personalized dosing! 18/n

Ok moving on to assumption #2a
Any evidence showing using AUC-based TDM improves efficacy?

Nope – no studies 😶
19/n

#2b
Any evidence showing AUC-based TDM is safer?

2 before-after studies → lower AKI with AUC vs. trough TDM 👍

aac.asm.org/content/61/12/…
aac.asm.org/content/62/2/e…
20/n

But all / many patients in trough arms dosed to target trough 15-20 → higher #vancomycin doses, higher exposure, longer duration in trough arm 👎
Strawman comparisons! 🙊21/n

In 2 other AUC vs trough observational studies, similar #vancomycin doses, exposures, durations &

NO difference in AKI 👊

bpspubs.onlinelibrary.wiley.com/doi/full/10.11…
accpjournals.onlinelibrary.wiley.com/doi/abs/10.100…
22/n

So.....these support what we knew → #vancomycin AKI related to ↑ dose, ↑ exposure (however you measure it!), ↑ duration, but don’t prove method of TDM per se improves safety 🤔 23/n

Without compelling evidence, AUC TDM
may be a path to ↑↑↑ resources without any clear benefit to patients or health systems 😕

We should not repeat mistake of implementing guidelines based on weak data!

2009 guidelines → AKI
2020 guidelines → Opportunity costs + ? ?
24/n

How can we improve outcomes in MRSA infections & limit #vancomycin-associated AKI?

1. ID consult for all S. aureus bacteremia
2. Source control
3. Don’t give #vancomycin to patients who don’t need it!!
4. Keep troughs < 15
5. RCTs like @snap_trial

Fini 🧵