MDRO Strategies

This presentation was presented at Apollo International Forum on Infection Control (AIFIC’ 2013), Chennai

The presentation is solely meant for Academic purpose

Tackling Multiple Drug

Resistant Organisms

Disclaimer: My Perspective

I have neither the

intellectual depth

nor

the luxury of the

remote academic view

that ID specialists or

clinical microbiologists

may have……

Intensive Care…

….is a “downstream” speciality…

…we face the consequences of

many therapeutic misadventures!!!

MDRO Definition “Micro-organisms (predominantly bacteria) that are resistant to one or more classes of anti-microbial agents

MRSA

VRE

ESBL Gram negatives “Pan-resistant” Acinetobacter KPC- Klebsiella pneumoniae

VIM-2 / NDM-1 Klebsiella pneumoniae Stenotrophomonas

Burkholderia VISA

VRSA…..

Assault of the Acronym

Considering where Antibiotics come from…

(not from the friendly neighbourhood drug dealer!)

…& the “Power of Bacteria”… (Historical & Numerical = Genetic)

Age: Bacteria 3,500,000,000 years (2000x) = 1yr

Eukaryotes 1,800,000,000 years (1000x)

Multi-cellulars 580,000,000 years (300x)

Australopithecus 4,000,000 years (2x)

Homo erectus <2,000,000 years (1) = 5.2 hrs

Antibiotic use 60 years (0.00003x) = 55 secs

Numbers: “ The number of E. Coli in the gut of each human being far exceed the number of people that now live or have ever inhabited the earth” !!! - Gould SJ; Life’s Grandeur

S. aureus

Penicillin

[1950s]

Penicillin-resistantS. aureus

Methicillin

[1970s]

Methicillin-resistantS. aureus (MRSA)

Vancomycin-resistant

enterococci (VRE)

Vancomycin

[1990s][1997]

Vancomycinintermediate-

resistantS. aureus

(VISA / GISA)[2002]

Vancomycin-

resistant

S. aureus

S. aureus

Penicillin

[1950s]

Penicillin-resistantS. aureus

Methicillin

[1970s]

Methicillin-resistantS. aureus (MRSA)

Vancomycin-resistant

enterococci (VRE)

Vancomycin

[1990s][1997]

Vancomycinintermediate-

resistantS. aureus

(VISA / GISA)[2002]

Vancomycin-

resistant

S. aureus

…antibiotic resistance is Inevitable !

Selection of Resistance

Antibiotic

exposure

Resistant strains: rare

No survival advantage

Resistant strains:

dominant

Use Promotes Resistance

JAC; DOI: 10.1093/ jac /dkg488

% Resistance

Usa

ge

(DD

D/1

000/

day

)

Urinary E. Coli Resistance vs. Antibiotic Use

Antibiotic use is a strong Correlate of Resistance

(but not the only factor)

Antibiotics & Resistance

Widespread use of antibiotics leads to the selection of antibiotic resistant strains

Archibald et al; ICARE

Clin Inf Dis 1997; 24: 211-15

Evidence:

Resistant organisms are more common amongst inpatients

ICUs (with highest use) have highest rates of resistance 0

5101520253035404550

% R

esis

tan

t O

rgan

ism

s

MRSA

MRSE

Pseud C

TZVRE

Inpatient

Outpatient

p <0.01 for all comparisons

Methicillin Resistant Staphylococcus Aureus

0

10

20

30

40

50

60

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Perc

en

t R

esis

tan

ce

Extended Spectrum Beta Lactamase (Klebsiella)

0

2

4

6

8

10

12

14

19891990

19911992

19931994

19951996

19971998

19992000

Per

cent

Res

ista

nce

7x

Quinolne Resistant Pseudomonas Aeroginosa

0

5

10

15

20

25

30

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Perc

en

t R

esis

tan

ce

3x

ESBL in the Developing World

Site Location %ESBL

Klebsiella E. coli

AIIMS, New Delhi1 Tertiary Hospital 80% -

Mathai 10 Tertiary Hosps. - >60%

KGMC, Lucknow2 Neonatal ICU 86% 64%

SMF, Chennai Nosocomial: ICU 84% 82%

SMF, Chennai Comm. Acquired: ICU 53% 44%

China, Shanghai3 University Hospital 51% 24%

Latin America4 SENTRY, Pneumonia 44% 29%

1: Ind J Med Res 2002;115:153-7 2: J Med Microb 2003; 52: 421-5 3: Zhou Yi Xue Za Zhi 2002;82:1476-9 4: Diag Mic Inf Dis 2002; 44: 301-11

Developing World ICUs International Nosocomial Infection Control

Consortium (INICC)

8 Countries; 55 ICUs; 46 Hospitals

NNIS (USA) INICC

1992-2004 2002-5

MRSA 59% 84%

ESBL 19% 55%

Quin. res Pseudo. 29% 59%

VRE 29% 5%

Ann Int Med 2006; 145: 582-91

Does ‘ESBL’ kill?

Case control study from Israel (198 patients)

Multivariate analysis;

ESBL remains independently associated

with

Mortality OR 3.6 (1.4 - 9.5) p <0.008

Delayed Rx OR 25.1 (10.5 - 60.2) p <0.001

LOS OR 1.56 p <0.001

Cost OR 1.57 p <0.003

Antimicrob Agents Chemo 2006; 50: 1257-62

VRE Implications: Mortality

Vancomycin Resistance is:

An independent predictor of death in

EC bacteremia (OR: 2.1; 95% CI:1.14-3.88)

Associated with J all cause mortality

(52% vs. 27% in vanco sensitive)

An independent predictor of infection

related death (OR: 5.2; 95% CI: 1.4-20.0)

Newer studies (? more effective Rx) show

no difference

Ann Int Med 2001; 135: 484-92

Diag Micr ID 2000; 36: 145-52

CID 2000; 30: 466-72

CID 2002; 34: 922-9

MRSA: Attributable Mortality

Clin Inf Dis 2003; 36:53–9

*Anes Clin N A 2004: 22 ; 405 - 35

Meta-analysis:

31 cohort studies

3963 patients

2/3 MSSA vs. 1/3 MRSA

Debated with VAP*

Odds of death 2x with MRSA bacteremia vs. MSSA

Risk Factors for MDRO

Age / co-morbidity

Severity of illness

ICU admission

Prior Antibiotic exposure

Prior colonization

Exposure to colonized/ infected patient

Invasive Device use

Demographic

Antibiotic / infection control

Procedural

CID 2001; 33: 939-46

Inf Cont Hosp Epi 2009; 30: 1180-5

Factors

Severity of illness….

Device use

Antibiotic use

Approach to Preventing MDRO

Avoid Devices

Minimise

device-associated

infection

Rational & conservative

antibiotic use

Early goal-directed

Rx

Cannot justify an “I’m not to blame” attitude

CDC: Strategies

Administrative Support

Education

Antibiotic Use

Active Surveillance Culture

Isolation Precautions

Environmental issues

Decolonization

Administrative Support

Education

Antibiotic Use

Active Surveillance Culture

Isolation Precautions

Environmental issues

Decolonization

Is our empirical strategy correct?

What is the value of a restricted formulary?

Practicality?

Culture vs. rapid detection?

Mupirocin & SDD?

Patterns of Antibiotic use in the ICU

Assumptions for use of Empirical Antibiotics

Infection J (attributable) mortality

Antibiotic treatment K mortality

Early administration of correct antibiotic is better than delayed administration

Adverse effects < the benefits of treatment

Attributable Mortality

Nosocomial Attributable

Infection mortality

All Nosocomial Infection 20% (2.8-44)

Blood-stream 25% (4.4-47.3)

Vent. Assoc. Pneumonia 25% (7.8-42)

Urinary Tract None Chest 2001; 120:2059–93

Definition:

Mortality caused by an infection (in excess of mortality in a similarly ill patient without infection).

Is death attributable to NI?

NI

Treat Don’t Treat

Is there a difference?

Yes No

NI kills Patient dies of

critical illness

Unethical to evaluate in this way!

Is death attributable to NI?

Cohort of ICU Patients

NI No NI

Rx No (appropriate) Rx

Severity of illness is a confounder;

May be adjusted by case-control studies

or Multivariate (regression) analysis

Ideally Propensity modeled

Definition:

Difference in death rates between

infected & uninfected patients after adjusting

for confounders

Effect of Appropriate Choice

Luna et al;

Chest 1997; 111: 676-85.

0

10

20

30

40

50

60

70

80

90

100

No Rx Correct Rx Incorrect

Ho

spit

al M

ort

alit

y (%

)

Effect of Appropriate Timing

0

10

20

30

40

50

60

70

80

90

100

Early Late

None Correct Incorrect

Ho

spit

al M

ort

alit

y (%

)

Luna et al;

Chest 1997; 111: 676-85.

Methodological Issues

CID 2007; 45: 329 - 37

Systematic review of 51 studies of “appropriate antibiotics” and mortality in bacteraemia

No distinction : Empirical vs. definitive

Only 16% (8) defined “appropriate” on the basis of in-vitro Cx + route + dose

Only 35% looked at attributable mortality

Only 21% (7) timed severity of illness measures; but 2 measured it at admission

33% (17) adjusted for septic shock

Only one study used Propensity scores

Lab confirmed

Suspect

infection

Avoiding Overuse: “De-escalation”

Initial Rx

Multiple drugs

wide-spectrum

De-escalate Rx

Chest 2002; 122:2183–2196.

Culture-based

de-escalation can

reduce resistance

Making De-escalation Possible De-escalation is seldom attempted

Recent study of VAP*

De-escalation only in 22% patients

Likely to occur if:

3-4 antibiotics were initially used

If adequate cover was initially provided

If major pathogen grew on culture

If culture method was more robust *Chest 2006; 129:1210–1218

Rational De-escalation?

Can we use sputum, BAL cytology & culture to make decisions?

Probability of VAP +ve -ve

BAL with bacteria >95% 25-50%

Sputum culture 95% 45%

BAL culture 72-95% 50%

No! JAMA. 2007;297:1583-93

Current Empirical Rx encourages overuse

Suspected VAP

Rx

VAP No VAP

Appropriate Inappropriate ?

Colonization by resistant bugs

& subsequent adverse outcome

65% 35%

No de-escalation

De-escalated

VAP: A Decision Analysis

A decision analysis based on available

data suggests:

Empirical Rx (using clinical diagnosis) :

66% mortality in the untreated patients

68% mortality in treated patients

Chest 1996; 110:1025-34

Reasons for unexpected death with Rx: Diagnostic certainty is low (0.23)

Uninfected pts colonize c resistant strains

Limitations: Old data; incomplete info

0 1 2 3 months

Restrict

Cycle

Heterogeneous use (Mixing)

Gorillamycin Supersporin Kingkongopenem

Cycling

Strategies to Reduce Resistance

Abx 2 Abx 3 Abx 1

Res

ista

nce

Abx 1

Presumes that a decline in resistance will occur with antibiotic cessation (ie, there is a ‘fitness cost’ to antibiotic resistance.)

Time

None of the evidence is “clean”

Mainly observational cohorts

Some studies are of a single scheduled change (not cycling)

Associated confounders:

Antibiotic restriction policies

Infection prevention strategies

Often evaulate nosocomial infection alone not colonisation rates

RCTs are needed

Cycling: The Evidence?

Gerding et al

AAC 1991; 35: 1284-90

Koleff et al

AJRCCM 1997; 156: 1040-8

Gruson et al

AJRCCM 2000; 162: 837-43

Raymond et al

CCM 2001; 29: 1101-8

Toltzis et al

Paediatrics 2002; 110: 707-11

Inefficacy of Cycling Restriction, Rotation, ?Rubbish…

Res

ista

nce

Abx 2 Abx 3 Abx 1 Abx 1

The rate of decline of resistance will depend on:

the cycling interval &

the fitness cost of antibiotic resistance

Proc Nat Ac Sci 1999; 96:1152-56

Mathematical Models

Model simulating hospital patients / NI

Bergstrom & Reluga:

Cycling and mixing are better than one drug

Non-optimal strategies of cycling or mixing degrade effect

Even in optimal conditions cycling is not much better than mixing

“Cycling is unlikely to reduce either the evolution or the spread of antibiotic resistance.” and “inappropriately optimized cycles may hinder resistance control”

Math Med & Biol 2005; 22: 187-208

PNAS 2004 ; 101: 13285–13290

Empirical Evidence

J Antimic Chemo 2006;

57: 1197–1204

44 month, single ICU evaluation of

4 strategies: Patient-specific antibiotic

Prioritized antibiotic (cycled)

Restricted antibiotic (cycled)

Maximal heterogeneity (mixed)

Antibiotic heterogeneity: Peterson Index

Homogeneity associated with

J Carbepenem Res. Acinetobacter : RR 15.5 (5.5–42.8)

J ESBL: RR 4.2 (1.9–9.3) and

J Enterococcus faecalis: RR 1.7 (1.1–2.9).

Preventive Strategies

Active Surveillance Cultures + Contact Precautions

Identify both colonized & infected patients early

Restrict spread by strict infection control processes

ASC + Precautions Best studied with MRSA; less with Gm Neg

ASC shows benefit in some; inconsistent

Mathematical modelling:

VRE: Culture (vs. none) K transmit ~40%

Culture + isolation K transmit 65%

MRSA: Routine cultures are ineffectual

ASC can reduce rates

Debated: When, how often, Which patients (all vs. at risk) Sites, Methods

Need RCT!

The RCT; You asked for it!

N E J Med 2011; 364: 1407-18

Cluster randomized study

10 intervention ICUs (surveillance Cx + barrier)

8 control ICUs (n= 5343 & 3705 respectively)

Surveillance cultures on all (reported to Rx grp)

MRSA+VRE colonization & infection unchanged

40.4+3.3 vs 35.6+3.7 / 1000 patient days (p 0.35)

Compliance with barrier precautions was suboptimal

The VA Initiative

2007 – 2010; 1.9 million admits ICU / non ICU 8.3 million patient days; “MRSA Bundle” initiated

Universal nasal surveill’nce: Chromgenic agar/ PCR

+ Contact precautions

K ICU MRSA ~62% (relative risk)

K Non-ICU MRSA 45%

“MRSA Bundle”

Universal nasal survl. Contact precautions Hand Hygiene “Infection control as everyone’s responsibility”

N E J Med 2011; 364: 1419-30

The Gist: STAR ICU vs. VA

While one can decry the deficiencies of observational cohorts; RCTs often underestimate benefits

The prime issue: methodology of surveillance

STAR: Routine Cx: Median positive ~ 5 days

VA: Chromgenic + PCR (“real time”)

Extrapolation to MDRO Gram negatives

Spot-tests for gram negative MDROs limited

Value of contact precautions is unsubstantiated

Decolonization:

I am not competent to discuss the merits of Mupirocin to prevent MRSA

But am tempted to introduce the concept of Selective Decontamination of the Digestive Tract (SDD)……

….interesting lessons!

The problem of

“Dutch” Ideas

Euthanasia

Legal drugs

Legal prostitution

“Going Dutch”

A function of living below sea level?

Methodology of SDD* *Selective Decontamination of the Digestive Tract

1. Oral application of antibiotic:

Tobra+ Amphotercin+ Polymixin

(Vancomycin added if MRSA is J)

2. NG administration

3. Systemic Antibiotics for 24-48 hr

CID 2006; 43: S70–4

1 & 2: “Selective”; anaerobes preserved

Reduce GI/ Oropharynx colonization

3: Prevents early colonization / VAP

Outcomes with SDD

D’amico

BMJ 1998; 316; 1275-85

Meta analysis

30 RCTs

5727 patients

Odds of VAP and death reduced;

VAP; OR 0.35 (0.29-0.41)

Death; OR 0.8 (0.69-0.93)

Death K from 30% to 24%

NNT to prevent one death = 16

Best effect in surgical ICU patients

Rx

Rx

So why not use SDD?

Resistance?

Fear of inducing bacterial resistance Not conclusively established

Resistance?

Will it work when rate of bacterial resistance is high in an ICU Has been used to combat outbreaks of resistant organisms; not in endemic high-resistance ICUs

AJRCCM 2001; 164: 382-88

AJRCCM 2001; 164: 338-9

AJRCCM 2002; 166: 1029-37

Intens Care Med 1999; 25: 1323-6

Rx

Rx

SDD in India?

Mathematical Modeling*

Maximum effect shown when: a. Colonization pr. with resistant bacteria is low

b. All patients at risk are given SDD

c. No external source of potential pathogens

Rx

Rx

* Bootsma et al

Intersci Conf Antimicrob Agents Chemother.

2003 Sep 14-17; 43:

abstract no. K-698.

None of the above can be met in our ICUs a. ESBL ~ 80%; MRSA ~50%; MDR

Pseudomonas~ 40%; MDR Acinetobacter

b. Not cost-effective to give SDD to everyone

c. Requires pre-emptive isolation (quarantine)

More Math Models!

The benefit of non-absorbable “antibiotic prophylaxis can only be substantial if patient-to-patient transmission has already been reduced to a subcritical level by barrier precautions”…..

“a firm theoretical argument against routine use of topical antimicrobial prophylaxis”

Tackling MDROs!

We are facing the consequences of decades of irresponsible use: So control processes will be expensive

Best option: ASC + Contact Precautions

Real-time surveillance need development

ICU Antibiotic use needs to be reduced

Prophylactic Abx only work after prevalence rates are reduced

Infection Control Processes remain supreme

“Antibiotics are a non-renewable resource”….

Laxminarayan & Brown 2001

…..use them wisely!