Changing MIC Breakpoints: Do Mechanisms Count

Changing MIC Breakpoints: 

Do Mechanisms Count 

Paul C. Schreckenberger, Ph.D., D(ABMM), F(AAM) 

Professor of Pathology 

Director, Clinical Microbiology Laboratory 

Loyola University Medical Center 

pschrecken@lumc.edu

YES

• The major issue is whether the new 

Enterobacteriaceae MIC breakpoints can 

predict clinical success and failure even 

without ancillary tests or whether laboratories 

need to continue to screen for, and confirm 

the presence or absence of ESBLs and 

Carbapenemases before issuing a 

susceptibility report 

3

CLSI Guidelines 

• CLSI has history of recommending testing for 

mechanisms in order to detect false susceptible 

results with in vitro testing 

Staph spp. Table 2C p. 69 M100-S21 

• Detection of oxacillin Resistance: Tests for mecA 

or PBP2a are the most accurate methods for 

prediction of resistance to oxacillin ….. Isolates 

that carry the mecA or PBP2a should be reported 

“oxacillin resistant” 

Staph spp. Table 2C p. 76,79 M100-S21 

• Inducible clindamycin rsistance can be detected by 

disk diffusion using the D-zone test ….. Report 

isolates with inducible clindamycin resistance as 

“clindamycin resistant” 

4

CLSI Guidelines 

Staph spp. Table 2C p. 70,78 M100-S21 

• An induced -lactamase test should be performed 

on staphylococcal isolates with penicillin MICs 

0.12 ug/ml or zone dia. 29 mm before reporting 

isolate as penicillin susceptible …. A positive - 

lactamase test predicts resistance to penicillin. 

5

CLSI Guidance on ESBL Testing 

• Will Tests for ESBLs and KPCs be needed with 

the new cephalosporin and carbapenem 

breakpoints for Enterobacteriaceae? 

CLSI says No. For patient management, 

tests for ESBLs and KPCs are not necessary 

If requested, tests for ESBLs and KPCs may 

be done for Infection Control purposes 

6

New CLSI Guidelines 2010 

• This decision was based on evaluation of 

pharmacokinetics-pharmacodynamics (PK-PD) 

properties and limited clinical trials 

• New interpretive criteria (breakpoints) establish 

for some (but not all) cephalosporins 

• Using new breakpoints treatment decisions can 

be based solely on MIC alone 

• “It’s all about the MIC stupid” 

7


• Problem with this approach 

Clinical studies show patients with 

“susceptible” MICs + ESBL fail therapy 

8

Clinical Studies 

• No randomized controlled trials have ever 

been performed that evaluated the use of 

various comparator antibiotics in treatment of 

serious infections due to ESBL-producing 

organisms 

• It is unlikely that such studies will ever be 

performed 

• Existing data comes only from retrospective 

studies 

9

Treatment of ESBL Positive 

Organisms with Cephalosporins 

MIC FAILURE DEATH 

8 100% (6/6) 33% (2/6) 

4 67% (2/3) 0% (0/3) 

2 33% (1/3) 0% (0/3) 

≤1 27% (3/11) 18% (2/11) 

(CLSI breakpoint 8 g/ml) 

Paterson, DL, et al. JCM 39: 2206 – 2212, 2001 

10



Clinical studies show patients with “susceptible” 

MICs + ESBL fail therapy 

Inoculum used in bmd (10 5 ) too low and may 

dilute out resistant subpopulations giving 

false susceptible results 

11

MICs in g/ml: SHV-3 producing Citrobacter freundii 

Inocul. 

CFU/ml 

Pitfalls of ESBL Testing 

Effects of Inoculum 

Cefotaxime Ceftazidime Aztreonam Cefepime 

5 x 10 5 2 1 0.5 0.5 


Thomson KS, Moland ES: Antimicrob Agents 

Chemother. 2001 Dec;45(12):3548-54 

12

MICs in g/ml: SHV-3 producing Citrobacter freundii 

Inocul. 

CFU/ml 

Pitfalls of ESBL Testing 

Effects of Inoculum 

Cefotaxime Ceftazidime Aztreonam Cefepime 

5 x 10 5 2 1 0.5 0.5 

5 x 10 7 256 32 32 >1024 


Thomson KS, Moland ES: Antimicrob Agents 

Chemother. 2001 Dec;45(12):3548-54 

13

Inoculum Effect 

• Animal studies also demonstrate an inoculum 

effect and adverse outcomes when 

cephalosporins are used to treat ESBLproducing 

organisms with MICs of 

cephalosporin in the susceptible range. 

14

Efficacy of different beta-lactams against an 

ESBL-producing K. pneumoniae in rat IAA Model 

• Abscess created by intraperitoneal placement gelatin 

capsule containing sterile rat cecal contents, killed B. 

fragilis, and 1:100 dilution of ESBL-K. pneumonia (ca. 

10 5 CFU) 

• Treatment begun by continuous infusion via left internal 

jugular vein 3-4 hours after placement of capsule and 

continued 3 days. 

• On day 2 and 3 blood samples were collected for 

determination of serum antibiotic levels 

• After 3 days animals were sacrificed, abdominal wall 

portion of abscess removed and quantitative culture 

performed 

Rice LB et al. Antimicrob Agents Chemother. 

1991 Jun;35(6):1243-4. 

15


ESBL-producing K. pneumoniae in rat IAA Model 

Table 1: MICs of various agents for K. pneumoniae 5657 

Antimicrobial Agent MIC at inoculum of: 

10 5 CFU/ml 10 7 CFU/ml 

Cefoperazone 2 256 

Sulbactam 32 

Cefoperazone-sulbactam (2:1) 0.5 256 

Cefotaxime 1 256 

Cefpirome 1 >256 

Ceftazidime >256 

Imipenem 0.5 16 


1991 Jun;35(6):1243-4. 

16


ESBL-producing K. pneumoniae in rat IAA model 

Table 2: Intra-abdominal abscess treatment outcomes 

Antibiotic No. 

rats 

Mean serum 

level ( g/ml) 

Log10 

CFU/g of 

abscess 

None 30 8.02 1.02 

Cefoperazone 11 13.5 4.72 7.41 0.74 a 

Cefoperazone-sulbactam 11 8.9 3.22 5.84 0.95 c 

Cefotaxime 18 17.7 8.42 7.26 1.02 a 

Cefpirome 11 28.3 2.06 7.80 1.18 a 

Ceftazidime 10 19.4 3.09 8.85 0.64 a 

Imipenem 19 7.1 2.08 4.99 0.97 c 

a P>0.05, c P



• Conclusions 

Extend spectrum cephalosporins may be less 

effective in treating serious infections due to 

ESBL producing gram-negative bacilli than 

standard susceptibility tests would imply 

In vitro studies indicated that the activity of 

these agents against K. pneumoniae 5657 

was highly inoculum dependent 


1991 Jun;35(6):1243-4. 

18



• Conclusions 

Dramatic improvement of in vivo efficacy of 

cefoperazone in presence of sulbactam was 

due at least in part to presence of 

-lactamase 

Avoid extended spectrum cephalosporins 

as single agents when treating serious 

infections with ESBL-producing 

organisms 


1991 Jun;35(6):1243-4. 

19



Clinical studies show patients with “susceptible” 

MICs + ESBL fail therapy 

Inoculum used in bmd (10 5 ) too low and may dilute 

out resistant subpopulations giving false susceptible 

results 

Different susceptibility testing methods give 

varying results 

20

CAP Survey D-C 2007 D-19 

C. freundii with PER-1 ESBL with MIC 16 g/ml 

Method/system 

tested (no.) 

Reported Cefepime 

MIC or zone in 

Susceptible category 

BD Phoenix (24) Not reported >16 

MicroScan (311) 15% >16 

Vitek (244) 97.2% 4 

Vitek 2 (230) 85.5% 2 

Final Critique Survey 2007 D-C 

College of American Pathologists 

Modal MIC 

Disk Diffusion (74) 34% 15.5 mm 

21

CAP Survey D-A 2009 D-05 

K. pneumoniae KPC with MIC ≥8 g/ml 

Carbabenem/ 

system tested 

(no.) 

No. Reports By Category 

Imipenem S I R Modal 

MIC g/ml 

BD Phoenix (24) 4 (17%) 7 13 >8 

MicroScan (538) 73 (14%) 217 248 8 

Vitek (154) 111 (72%) 10 33 ≤4 

Vitek 2 (422) 44 (10%) 6 372 2 

Final Critique Survey 2009 D-A 

College of American Pathologists 

22



Clinical studies show patients with “susceptible” MICs + ESBL 

fail therapy 

Inoculum used in bmd (10 5 ) too low and may dilute out resistant 

subpopulations giving false susceptible results 

Different susceptibility testing methods give varying results 

MIC results are not reproducible and can very up to 

>3 dilutions upon repeat testing 

23

CLSI Acceptable Limits ( g/mL) for QC 

Strains Used to Monitor Accuracy 

Antimicrobial E. Coli ATCC 25922 No. of Doubling 

Dilutions Allowed 

Cefazolin 1-4 3 

Cefotaxime 0.03-0.12 3 

Ceftriaxone 0.03-0.12 3 

Ceftazidime 0.06-0.5 4 

Aztreonam 0.06-0.25 3 

Cefepime 0.015-0.12 8 

CLSI M100-S20. Table 4 

24

MIC Reproducibility 

• In this imperfect reality an ESBL producer 

with a cefotaxime MIC of 1 g/mL (probably 

responsive to cefotaxime in vivo) cannot be 

reliably distinguished from one with an MIC of 

4 g/mL (probably not responsive) and it is 

simpler and safer to follow the precautionary 

approach of seeking the ESBL and, if found, 

reporting the isolate as resistant 

- David Livermore, 20 th European Congress of 

Clinical Microbiology and Infectious Diseases, 

Vienna, Austria, 10-13 April, 2010. 

25




fail therapy 




MIC results are not reproducible and can very up to 3 dilutions 

upon repeat testing 

Some cephalosporin breakpoints were not lowered 

enough and others not lowered at all 

26

Enterobacteriaceae 

Revised Breakpoints (MIC g.ml) 

Agent CLSI M100-S19 

(2009) 

CLSI M100-S20 

(2010) 

Susc Int Res Susc Int Res 

Cefazolin 8 16 32 1 2 4 

Cefotaxime 8 16-32 64 1 2 4 

Ceftizoxime 8 16-32 64 1 2 4 

Ceftriaxone 8 16-32 64 1 2 4 

Ceftazidime 8 16 32 4 8 16 

Aztreonam 8 16 32 4 8 16 

CLSI M100-S20. Table 2A 

27

Enterobacteriaceae – Evaluated 

but NOT Revised Breakpoints 

Agent CLSI M100-S20 (2010) 

Cefuroxime 

(parenteral) 

Susc Int Res 

8 16 32 

Cefepime 8 16 32 

Cefotetan 16 32 64 

Cefoxitin 8 16 32 

CLSI M100-S20. Table 2A 

28

Cephaloporin Breakpoints Following 

Recent EUCAST and CLSI Revisions 

Group Cefuroxime 

S /R> 

Cefotaxime 

S /R> 

Ceftriaxone 

S /R> 

Ceftazidime 

S /R> 

Cefepime 

S /R> 

CLSI 8/8 1/2 1/2 4/8 8/16 

EUCAST 8/8 1/2 1/2 1/8 a 1/8 a 

EUCAST 

PK/PD 

4/8 1/2 1/2 4/8 4/8 

a Ceftazidime and cefepime S breakpoints were adjusted from 4 to 1 ug/ml to 

ensure that Enterobacteriaceae with clinically important ESBLs were not 

reported as susceptible 

Kahlmeter G Clin Microbiol Infect. 2008 Jan;14 Suppl 

1:169-74. Review 

29

Enterobacteriaceae epidemiologic 

cut-off values (wild type X g/mL) 

E. coli K. pneumoniae K. oxytoca P. mirabilis 

Cefuroxime 8 8 8 4 

Cefotaxime 0.25 0.12 0.12 0.06 

Ceftriaxone 0.25 0.12 0.12 0.06 

Ceftaxidime 0.5 0.5 0.5 0.12 

Cefepime 0.12 0.12 0.12 0.12 

Kahlmeter G Clin Microbiol Infect. 2008 Jan;14 Suppl 

1:169-74. Review 

30




fail therapy 




MIC results are not reproducible and can very up to 3 dilutions 

upon repeat testing 

Some cephalosporin breakpoints were not lowered enough and 

others not lowered at all 

ESBL enzymes are substrate specific, may miss 

ESBL resistance if not testing the preferred substrate 

(eg. CTX-M) 

31

E. coli with CTX-M ESBL 

33

Why labs should continue to perform 

ESBL and KPC testing on all isolates 

•MIC results are inoculum dependent. 

•MIC results are not reproducible and can vary 

by >3 dilutions 

•Some ESBLs are inducible, ie. MIC is low initially 

but will test resistant after exposure to 

antibiotic. Presents a problem especially for 

Rapid Detection Methods 

•Knowing mechanism of resistance is important. 

It allows us to modify our reports and 

implement prompt infection control strategies 

34

Why labs should continue to perform 

ESBL and KPC testing on all isolates 

•Combining the lowering of MIC 

breakpoints with ESBL and KPC confirmation 

testing gives the laboratory the best chance 

of getting the result right for the patient 

35

Are the New CLSI Breakpoints 

the Correct Breakpoints?

Something to Think About 

• The MIC is only one piece of information that is 

useful in deciding if an antibiotic is going to be 

successful in treating an infection. 

• There will never be a single breakpoint that will 

predict susceptibility or resistance to a drug for 

all patients, at all sites of infection, and under all 

conditions based on the MIC 

37


• Breakpoints can vary by infection type (e.g., 

cystitis vs. sepsis vs. meningitis, etc.). 

• Breakpoints can vary by the method of antibiotic 

administration (e.g., oral vs. IV administration) 

or dosage given (customary vs. highest dose). 

• The FDA and CLSI breakpoints are based only 

on the drug level achieved in blood using the 

customary dose 

38


• In fact, the antibiotic reports issued by the 

laboratory are most often not for infections 

occurring in the blood (LUMC Feb ’11 87/813) 

• Drug concentrations at sites of infection 

outside blood stream may be higher or lower 

than those in blood, rendering the breakpoint 

less useful in these sites. 

39


• Antibiotics are often renally excreted and the 

renal function of the patient will determine 

how long the antibiotic will remain in the 

blood and how much is concentrated in other 

sites such as the bladder 

• In addition, all antibiotics are partially protein 

bound, so the albumin level of the patient 

affects the amount of free drug circulating. 

40

Time to Start Thinking Outside 

the Box – Eliminate Breakpoints 

• I proposed a new approach for interpreting 

the results of AST, customized for each 

patient based on a number of factors 

41

Personalized Antibiotic Report 

• Customized for each patient based on: 

Organism ID 

MIC 

Site of infection 

Drug concentration 

at site of infection 

Renal function 

Protein level 

Method of 

administration 

Dosing 

convenience 

Potential for D-D 

interactions 

Patient allergies 

Cost 

42


• To achieve this goal, sophisticated computer 

software would have to be developed that 

would customize reports for each patient by 

mining the EMR and using data on creatinine 

clearance, albumin concentration, site of 

infection, type of infection, culture results, and 

mode of antibiotic administration, to provide a 

personalized antibiotic report that would 

specify the optimal antibiotic(s) to treat the 

infection 

43


• Cost features, such as the need to perform 

therapeutic drug monitoring, drug-to-drug 

interactions, and in-patient or home therapy 

usage could also be incorporated into the 

computer algorithm 

We a teenager or a guy 

named Steve to step up 

and develop a computer 

application for this purpose 

44

Is it time to move past the 

concept of the universal 

breakpoint? 

Let’s Think About It

Are the New CLSI Breakpoints 

the Correct Breakpoints? 

“There ain’t no answer. There 

ain’t going to be an answer. 

There never has been an 

answer. That’s the answer” 

- Gertrude Stein

Changing MIC Breakpoints: Do Mechanisms Count

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?