antibiotics: the many vs. the few - critical care canada

Antibiotics: The many vs. the few

Brian H Cuthbertson Chief of Critical Care Medicine

Sunnybrook Health Sciences Centre Professor, Critical Care Medicine

University of Toronto Toronto Canada

There are only three responses to a new paper!

• “This work agrees with my bias” • “I am smart enough to know that without

the need for a trial”

I already knew that!

• This disagrees with my personnel bias • No trial is going to convince me on this one

regardless of the size

I don’t believe it!

• I don’t believe it (but clever people are in the room!)

I have concerns with the generalisibility of the result!

Antibiotics: The many vs. the few

The example – SDD!

SDD- The background

• Hospital acquired infections significant problem in all hospital

• 20-50% critically ill suffer from HAIs

• Traditionally, HAI in critical illness were from Gram negative enteric bacteria

• This has changed with the rise of MRSA

SDD- History

• First description in intensive care in 1983

• Flurry of publications from late 80s and 90s

• Large RCTs published in last 10 years

• Used in some areas of NW Europe (Holland)

• Not widely adopted elsewhere in the world

• Not used in ICU practice in North America

What actually is SDD?

• ‘Selective decontamination’ NOT ‘sterilisation’

• Target enteric aerobic Gram negatives

• Gastric overgrowth and subsequent VAP

• Bacterial translocation and metastatic sepsis

• Attempts to not target anaerobes and Gram positives

• Beneficial bowel flora “Good bacteria”

Which antibiotics, where?

• Topical oral/enteral, non-absorbable antibiotics for duration of ICU admission

– Polymyxin B

– Tobramycin

– Amphotericin B

• IV cefotaxime (or ciprofloxacin) for 4 days or until surveillance cultures demonstrate GIT decontamination

It’s not new

Extensively Studied

• At least 60 clinical trials identified on Medline

So….

What is the problem with SDD?

Why don’t we use it…

“There is no evidence”

Why don’t we use it…

Does SDD benefit the individual to who it is delivered (the few)?

36 RCT and 11 meta-analyses

Effects of SDD on Survival

De Smet, Bonten et al, NEJM, 2009

• 36 trials, 6914 patients [17 more excluded]

• Topical plus systemic 17 trials, 4295 patients

• Omitted the De Smet study

OR 0.75, 95% CI 0.65 to 0.87

SDD and mortality

SDD and Resp. tract infection

0.28, 95% CI 0.20 to 0.38

Our meta-analysis

0.73, 95% CI 0.65 to 0.81

Does SDD benefit the individual to who it is delivered (the few)?

Yes!

Question…

Question…

What further evidence do we need?

Do we actually believe it benefits

the few?

Question…

Do we believe it benefits the few?

“Overall, SDD benefits the patients to whom

it is delivered”

Median= 6, IQR=5-7, Importance= 6

Do we believe it benefits the few?

“SDD reduces VAP”

Median=7, IQR= 5-8, Importance= 7

Do we believe it benefits the few?

Median= 4, IQR= 3-5, Importance= 7

“There is no mortality benefit associated

with SDD”

Do we believe it benefits the few?

“SDD is not on my unit’s list of clinical priorities”

Median= 8, IQR= 7-9, Importance= 6

Do we believe it benefits the few? “The SDD evidence base is not generalizable

to my country ”

Mean= 5, IQR= 5-7, Importance= 6

• “This work agrees with my bias” • “I am smart enough to know that without

the need for a trial”

I already knew that!

• This disagrees with my personnel bias • No trial is going to convince me on this one

regardless of the size

I don’t believe it!

• I don’t believe it (but clever people are in the room!)

I have concerns with the generalisibility of the result!

Do we believe it benefits the few?

“It is ethically acceptable to conduct further RCTs

evaluating the effectiveness of SDD”

Median=6 , IQR=5-7

01

02

03

0

Pe

rce

nt

1 2 3 4 5 6 7

So do we believe it benefits the few?

No!

Question…

“There's only two things I hate in this world. People who are intolerant of

other people's cultures and the Dutch”

Question

Does SDD benefit the the ecology of the unit /

hospital (the many)?

Why don’t we use it…

We’re worried about anti-microbial resistance

(the many)

• Colonisation with resistant gram negatives in 16% of SDD pts vs 26% of controls (p = 0.001)

• Colonisation with VRE in 1% of each • No MRSA

Cluster RCT of SDD in NEJM

• Background: SDD and SOD infection prevention measures in ICU but reported effects on patient outcome conflicting.

• Design: non-blinded RCT cross over design in 13 ICUs Netherlands

• Patients: 5939 ICU patients expected ventilation ≥ 48 hours and LOS ≥ 72 hours

• Interventions: Each ICU randomized to SDD, SOD, and standard care over 6 month period.

• Outcomes: 28 day mortality primary endpoint

De Smet, Bonten et al, NEJM, 2009

• Low rates of multi-resistant infections (5%) • No MRSA • Lower antibiotic use in SDD units

10% 6% 12%

“Widespread use of SDD and SOD

is justified”

“Widespread use of SDD and SOD

in intensive care units

with low levels of antibiotic resistance

is justified”

Meta-analysis of SDD and antibiotic resistance- VRE

Meta-analysis of SDD and antibiotic resistance- MRSA

Meta-analysis of SDD and antibiotic resistance

Aminoglycosides Polymixins

Fluoroquinolones Cephalosporins

Does SDD increase or reduce antibiotic use?

So do we believe it harms the many?

Question…

Do we believe it harms the many?

“SDD increases antibiotic resistance”

Median= 6 , IQR= 5-7, Importance= 9

Do we believe it harms the many?

“SDD would increase C. Difficile infections”

Median= 5 , IQR= 5-5, Importance= 8

Do we believe it harms the many?

“My concerns about antibiotic resistance limit my

willingness to participate in future RCT’s of SDD”

Mean= 4, IQR= 3-6

Do we believe it harms the many? “I would be more likely to participate in an RCT if it included

pre, during and post-trial monitoring

of antibiotic resistance in all patients in the RCT ”

Mean= 9, IQR= 9-9

“I am opposed to SDD”

ICU physicians

Micro / ID physicians

ICU pharmacists

ICU clinical leads

“Overall the risks outweigh the benefits”

ICU physicians

Micro / ID physicians

ICU pharmacists

ICU clinical leads

Do we believe it harms the many?

“There are conflicting opinions between

microbiologists and Intensive care clinicians”

Mean= 7, IQR= 6-9, Importance= 7

Do we believe it harms the many?

• We interviewed world leading trialists

• They generally believed SDD was beneficial

• But still wanted further trials!!

Conclusions

It’s a mess!

Conclusions

We have equipoise but maybe shouldn’t!

Programme of research

• A multi-centre, cluster RCT (SuDDICU-RCT)

• An contemporaneous ecological study (e-SuDDICU)

• A concurrent prospective economic evaluation of SDD (SuDDICU-CEA)

Phase 3 research design

SuDDICU design

We hypothesise that-

• SDD will reduce hospital mortality

• SDD will be cost-effective

• SDD will not harm the ecology of the ICU

• SDD will not increase important antibiotic resistance patterns

Hypotheses

Primary Outcomes

• The primary effectiveness outcome:- Hospital mortality

• The primary ecology outcome:- The difference in the incidence of antibiotic resistant organisms per 1000 patient admissions

• Primary CEA outcome- Cost-effectiveness over lifetime horizon