18. antibiotic use in the ic ul

Antibiotics in the ICU Antibiotics in the ICU

Ram E. Rajagopalan,MBBS, AB (Int Med) AB (Crit Care)

Department of Critical Care Medicine

SUNDARAM MEDICAL FOUNDATIONChennai

Choice, Resistance, Pharmacokinetics & Pharmacodynamics

Nosocomial InfectionBetween 30-40% ICU patients have or develop infectionAbout ½ of these are nosocomial

Prevalence is very variable between different kinds of units

Bacteriologicallyconfirmed

(true positive)

“Anticipated” infection

Clinicallysuspectedinfection

Antibiotic use in the ICU

Empirical Rx

Definitive Rx

Prophylactic Rx

Indications for antibiotics ~ 60% of patients in a general ICU receive antibiotics

Bergmans et alJAC 1997; 39: 527-35

Inci

sion

Tim

e

Time of Prophylactic dose

Rat

e of

Wou

nd In

fect

ion

Antibiotic Prophylaxis

Only for “clean-contaminated” & some “clean” wounds Appropriate cover; No value for additional doses

One dose..

..one hour before....incision

SURGICAL WOUND

Antibiotic Use & Resistance

Habarth et al.Circulation 2000; 101: 2916-21

….but prolonged ABP was an independent risk factor for bacterial resistance (AOR 1.6; 95%CI: 1.1-2.6)

2641 CABG patients; 4-yr. observ. cohortEither < or > 48 hr antibiotic prophylaxis

Systemic Antibiotic Use

0

5

10

15

20

25

30

35

0 5 10 15 20

Days of ventilation

% w

ith V

AP

No Abx

Prior Abx

Resistant Ps. aeroginosa; Acinetobacter

19%

65%

ARRD 1989; 139:877–84.

Ventilator AssociatedPneumonia

Policy on Prophylaxis

Limit antibiotic prophylaxis to surgicalwounds; ideally one pre-op. dose

Avoid forPreventing pneumonia in intubated ptsPreventing bacteremia from CV linesPreventing UTI from a Foley’s catheter

#1Prevention Cure

Empirical AntibioticsSuspected InfectionEmpirical Treatment

Infection:Subsequentlyconfirmed bybacteriology

Infection:not confirmable(false negative)

No infection(SIRS)

? Over treatment

25-40%60-75%

ICEM Yearbook 2003; 218-228.

Assumptions for use of Empirical AntibioticsNosocomial infection increases (attributable) mortality

Antibiotic treatment reduces mortality

Early administration of correct antibiotic improves survival

Adverse effects are less than the benefits of treatment

Is death attributable to NI?

NI

Treat Don’t Treat

Is there a difference?

Yes No

NI kills Patient dies ofcritical illness

RIPRIP

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

Does Rx improve survival?

35 Clinical Trials identified

None placebo controlledOnly 2 double-blind

No “Gold Standard” antibioticMainly comparing antibiotic regimensMainly “equivalence” trials

NO DIFFERENCE BETWEEN AGENTS

Where are theEmperors clothes?

ICEM Yearbook 2003; 219-28.

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 studiesor Multivariate (regression) analysis

About 25% attributable mortalityRIPRIP

Definition:Difference in death rates between infected & uninfected patients after adjusting for confounders

Attributable Mortality

Nosocomial AttributableInfection mortality

All Nosocomial Infection 20% (2.8-44)Blood-stream 25% (4.4-47.3)Vent. Assoc. Pneumonia 25% (7.8-42)Urinary Tract NoneChest 2001; 120:2059–93

RIPRIP

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

Hos

pita

l Mor

talit

y (%

)

Effect of Appropriate Timing

0

10

20

30

40

50

60

70

80

90

100

Early Late

None Correct Incorrect

Hos

pita

l Mor

talit

y (%

)

Luna et al; Chest 1997; 111: 676-85.

Timing and Mortality

Iregui et al Chest 2002; 122:262–8

No Delay12.5 + 4.2 h

Delay28.6 + 5.8 h

Tim

e fr

om D

x of

VA

P (h

)

20

40

n=107

Delay increased hospital mortality 2.5x& VAP attributable mortality 3.5x

RIPRIP

Design Flaws

VAPRx

Culture + Culture -

Appropriate Inappropriate ?

Colonization by resistant bugs& subsequent adverse outcome

25% 75%

Making the Choice

Appropriately timed, appropriately chosen antibiotic can save lives…….so…. ….which antibiotic(s)?

Focused Rx with narrow spectrum (or)

Gorillamycin + Supersporin +

King-Kongopenem?

CTZ Cipro Oflox Amik Cef Sul Pip Taz Imepen

Pseudomonas (46)

30%

(13/44)

24%

(11/46)

23%

(10/43)

36%

(16/44)

46%

(21/46)

60%

(26/44)

64%

(7/11)

Klebsiella ESBL (38)

0%

(0/38)

5%

(2/38)

26%

(10/38)

35%

(13/37)

32%

(12/38)

37%

(14/38)

80%

(8/10)

E Coli ESBL (36)

0% 3%

(1/36)

- 58%

(21/36)

53%

(19/36)

67%

(24/26)

82%

(9/11)

Acinetobacter (15)

8%

(1/12)

31%

(4/13)

73%

(8/11)

75%

(9/12)

55%

(6/11)

36%

(4/11)

100%

(3/3)

Proteus (18)

72%

(11/15)

11%

(2/18)

11%

(2/18)

67%

(12/18)

100%

(15/15)

87%

(13/15)

100%

(2/2)

Non-ESBL (K+Ec) (15)

100%

(15/15)

93%

(14/15)

- 100%

(15/15)

100%

(10/10)

100%

(10/10)

100%

(2/2)

Antibiogram ™: SMF ICU Nosocomial Pathogens,

Antibiotics & Resistance

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

Archibald et al; ICAREClin 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

tant

Org

anis

ms

MRSAMRSE

Pseud C

TZVRE

Inpatient

Outpatient

p <0.01 for all comparisons

“De-escalation”

Chest 2002; 122:2183–2196.

De-escalate Rx

Lab confirmed

Initial Rxwide-spectrum

Suspectinfection

Culture-basedde-escalation can reduce resistance

Policy on Empirical Rx

a. Empirical Rx can be justified in ICUb. Abx. choice based on local patterns c. Initial Rx should be wide-spectrum d. De-escalate based on culture result

As choice is based on local susceptibility& de-escalation needs culture resultsCULTURE prior to antibiotics is MANDATED!

#2

Policy for Definitive Therapy

Critically ill patients have microbiological colonisation of normally sterile sites

Treat true infection

Aviod antibiotics for colonised patients

!#3

Microbe Culture in VAP

Prob

abili

ty

1 2 3 4 5 6

VAP

Log CFUs

No VAP

CCM 2003; 31: 2544 – 51.

Routine culture of sputumDoes not differentiate infection vs. colonization

Quantitative or semi-quantitative culturesare needed

Responsible Use

1. Identify the site of infection

2. Know your bacteria

3. Maximise efficacy / minimise toxicity Appropriate dose ; PK-PD understood Drug level monitoring Duration

#4

Identify Site

SMF ICU 2001

05

1015202530354045

Pneumonia Bacteremia Urinary Tract

Pseudomonas Klebsiella E. Coli

Proteus Staph aureus Enterococcus

CTZ Cipro Oflox Amik Cef Sul Pip Taz Imepen

Pseudomonas (46)

30%

(13/44)

24%

(11/46)

23%

(10/43)

36%

(16/44)

46%

(21/46)

60%

(26/44)

64%

(7/11)

Klebsiella ESBL (38)

0%

(0/38)

5%

(2/38)

26%

(10/38)

35%

(13/37)

32%

(12/38)

37%

(14/38)

80%

(8/10)

E Coli ESBL (36)

0% 3%

(1/36)

- 58%

(21/36)

53%

(19/36)

67%

(24/26)

82%

(9/11)

Acinetobacter (15)

8%

(1/12)

31%

(4/13)

73%

(8/11)

75%

(9/12)

55%

(6/11)

36%

(4/11)

100%

(3/3)

Proteus (18)

72%

(11/15)

11%

(2/18)

11%

(2/18)

67%

(12/18)

100%

(15/15)

87%

(13/15)

100%

(2/2)

Non-ESBL (K+Ec) (15)

100%

(15/15)

93%

(14/15)

- 100%

(15/15)

100%

(10/10)

100%

(10/10)

100%

(2/2)

Antibiogram ™: SMF ICU Nosocomial Pathogens

ExtendedSpectrumBeta Lactamase(Klebsiella)

2002

7x

The ESBL “Plague” Plasmid-mediated resistanceVery high frequency in Indian Hospitals

MYSTIC study; Mathai et alDiag Mic Inf Dis 2002; 44: 367-77.

~70% of E. coli / Klebsiella all over India >50% in Anna Nagar ‘community’ >80% in SMF ICU

Cross-species transmission amongst enterobactereaciaeWide resistance to all beta-lactams including cephalosporins !

ESBL in the Developing World

Site Location %ESBLKlebsiella 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

What provokes ESBL Resistant Organisms?

Chest 2001; 119: 397S - 404SChest 2001; 119: 391S - 396SAJRCCM 2000; 162:1610-16J Hosp Infect 2003 ; 53: 39-45JAMA 1999; 281: 517-23

Are there identifiable risk groups?Prior 3rd Gen Ceph. Poor Functional Status

High APACHE score>5 days in ICUDecubitus ulcersInstrumentation

Community wide use of ciprofloxacin or co-trimoxazole!

Does ‘ESBL’ kill?Case control study from Israel (198 patients)

Multivariate analysis;

ESBL remains independently associated withMortality OR 3.6 (1.4 - 9.5) p <0.008Delayed Rx OR 25.1 (10.5 - 60.2) p <0.001LOS OR 1.56 p <0.001Cost OR 1.57 p <0.003

RIPRIP

Antimicrob Agents Chemo 2006; 50: 1257-62

Treating ‘ESBL’:

Chest 2001; 119: 391S - 396S.Diag Mic Inf Dis 2002; 44: 367-77.

Carbepenems are drugs of choice

Avoidance of all 3rd gen. CephalosporinsTraditional choices for ‘serious’ gram negative infectionsIn-vitro sensitivity may be spurious

4th Generation Ceph. may be effective

B-lactam / lactamase inhibitors; recommended but have high rates of resistance

Aminoglycoside, quinolone resistance common

Minimizing ESBL ResistanceRestriction of Ceftazidime alone is not effective

Hospital- wide restriction of 3rd generation cephalosporins

Use of B-lactam/ lactamase inhibitor for empirical therapy

Good infection control measuresJAMA 1998; 280: 1233-7

MDR AcinetobacterIncreasing prevalence and outbreaks

Gram negative “cocco-bacillus”

Ventilator Associated PneumoniaBacteremiaSecondary meningitis / ventriculitisSkin infectionsEndocarditisCAPD-peritonitis

TEM1 / CARB mutationESBLOxa- beta lactamaseCarbepenemaseA/G, quinolone resist

Environmental Sources of Acinetobacter

Bed rails Bedside tables Ventilators Infusion pumps Mattresses Pillows Patient monitors

X-ray view boxes Curtain rails Curtains Equipment carts Sinks Ventilator circuits Floor mops

Infect Control Hosp Epidemiol.

2003;24: 284-95

Risk Factors for Acinetobacter

Ventilator Assoc. Pneumonia

Age Chronic lung Dx Immunosupress. Prior antibiotic Rx Invasive Device Prolonged ICU stay

Bloodstream

Infection

Respiratory failure Prior Antibiotics Invasive device________________________________________________

72% of BSI has a lung

sourceCID; 2001; 33: 939-46

Impact of Acinetobacter in the ICUOutcome Group Any infection Pneumonia

MortalityCases 58% 70%

Controls 15% 17%

Attributable mortality

43% 53%

Risk ratio for death

4.0

(CI951.9-8.3)

4.0

(CI951.6-10.2)

Length of stay (median)

Cases 23 days 23 days

Controls 10 days 10 days

Excess LOS 13 days 13 days

Crit Care Med 1999;27:1794-1799. 48 patients 1:1 case control study

Impact of A. baumanii Resistance

Outcome Group All Imipenem (R)

MortalityCases 40% 44%

Controls 28% 24%

Attributable mortality

12%P=NS

20%P=NS

Length of ICU stay (median)

Cases 35 days

Controls 37 days

Excess ICU LOS-2 days

P=NS

Crit Care Med 2003;10:2478-2482. Historical cohort study; 60 patients with A. baumanii VAP matched 1:1 to Controls

Acinetobacter: Rx Options

Carbepenem (if no resistance)Colistin / Polymixins (aerosol?)Sulbactam (ampi-sulb; cefperazone-sulb)

66

33

21

86

37

47 47

0

20

40

60

80

100

Amp/sulb Pip Cefotaxime Imipenem Cipro Gent TMP/SMX

MDR, A. baumaniIs resistant to chlorhexidine

J Hosp Infect 2002;51:106-113.

TSN Database

QuinolneResistantPseudomonas Aeroginosa

0

5

10

15

20

25

30

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Per

cen

t R

esis

tan

ce

2002

3x

Carbepenem Resistant P. aeruginosaIncreasing prevalenceAmp C mechanism; beta lactams induce

Carbepenem resistance:Mutational porin channel lossAcquired metalo-beta-lactamases (integrin)

Treat: Colistin, Aminoglycoside, Quinolone, 4th G Cephalosporin

Prevent: ? multi-agent Rx of all ICU P. aeruginosa

MethicillinResistantStaphylococcusAureus

2002

How About MRSA?

Quite common; VAP, bacteremia

Risk factors: Age, severity of illness,prior antibiotic use, instrumentation*

Attributable mortality (>MSSA) debatable

Prefer Vancomycin for Rx even in VAP or severe CAP; equivalent efficacy@ to linezolid

*Ann Intern Med. 2002; 136: 834-44@J Anti Chemo 2003; 51(S2): ii27-35

Pharmacokinetics: VdPharmacokinetics is altered and variable in critically ill

Volume of distribution (Vd) can change significantly in any individual with changing fluid balance

Vd is vastly different between patients

Implication: Regardless of organ dysfunction most drugs need to be “loaded” at normal or higher doses.

Drug level measurement would be ideal

Surgery 1998; 124: 73-8J. Chemother 1995; 7: 45-9Intens Care Med 2001; 27: 363-70Ant Ag & Chemo 2002; 46:1557-60Intens Care Med 2004; 30: 2145–56

Clearance: Renal FunctionIncreased Vd : Due to fluid retention Vd and Renal failure = clearance

Calculated Creatinine clearance / GFR** MDRD or Cockraft-Gault method are poor

: in acutely ill with changing renal function: severe muscle wastingMeasured values ideal

CRRT similar to GFR ~35 ml/min* CVVHD clears < 50 Da molecular weightCVVF clears larger molecules (>1100 Da)

*Intens Care Med 2004; 30: 2145–56**N E J Med 2006; 354: 2473-83

Clearance: Liver FunctionLiver function does not affect most antibiotic clearances significantly

Hypoalbuminemia may increasefree antibiotic levels, but again the effect is modest and does not affect treatment

Chloramphenicol, clarithro, erythro, clinda,

metronidazole, nafcillin, tetracycline… may need some dose adjustmentIntens Care Med 2004; 30: 2145–56

Pharmacodynamics

MBCMIC

Bac

teria

l Cou

nt a

t site

Ant

ibio

tic C

onc.

Post Ab Effect

Pharmacodynamics:Defines the Interaction of the drugs and the microbes

Inf Dis Clin NA 2004; 18: 451-65

Concentration Dependent

MBCMIC

Bac

teria

l Cou

nt a

t site

Ant

ibio

tic C

onc.

Peak

AUC

Indicators of efficacy:Peak / MIC ratioAUC / MIC ratio

Inf Dis Clin NA 2004; 18: 451-65

Agents:QuinolonesAmphotercinAminoglycosidesMetronidazole

Time Dependent

MBCMIC

Bac

teria

l Cou

nt a

t site

Ant

ibio

tic C

onc. Time > MIC

Inf Dis Clin NA 2004; 18: 451-65

Indicators of efficacy:Time above MICPreferably > 50%

Agents:PenicillinsCephalosporinsCarbepenemsGlycopeptidesMacrolidesLinezolid

Value of PK / PDBetter Drug Dosing: Once-a-day aminoglycoside therapy

Better interpretation of antibiotic levels-improved responses

Population studies allows better identification of MIC breakpoints

Better dosing to minimize emergence of resistance

Duration of TreatmentWhat is the optimal duration of antibiotic

administration?

Guaranteecure

Minimize costMinimize resistance

Should resistant strains be subjected to longer treatment courses?

JAMA 2003; 290: 258-98

0 1 2 3 months

0 1 2 3 months

0 1 2 3 months

Restrict

Rotate

Heterogeneous use

Gorillamycin Supersporin Kingkongopenem

Rotation

Strategies to Reduce Resistance

Abx 2 Abx 3 Abx 1

Res

ista

nce

Abx 1

Presumed that the decline in resistance with antibiotic cessation would be at least as fast as the emergence of resistance.

Time

None of the evidence is “clean” Mainly observational cohorts Some studies are of a single scheduled

change (not rotation) Associated confounders:

Antibiotic restriction policies Infection prevention strategies

Often evaulate nosocomial infection alone not colonisation rates

RCTs are needed

Rotation: The Evidence?

Gerding et alAAC 1991; 35: 1284-90

Koleff et alAJRCCM 1997; 156: 1040-8

Gruson et alAJRCCM 2000; 162: 837-43

Raymond et alCCM 2001; 29: 1101-8

Toltzis et alPaediatrics 2002; 110: 707-11

Strategies to Reduce ResistanceRestriction, Rotation, ?Rubbish…

Res

ista

nce

Abx 2 Abx 3 Abx 1Abx 1

Resistance emerges fast (with antibiotic use) but…… declines slowly (with cessation of use)*

Antibiotic cycling is inferior to “heterogeneous” use**

*Austin et alProc Nat Ac Sci 1999; 96:1152-56

**Bonhoeffer et alProc Nat Ac Sci 1997; 94: 12106-11

The Power of BacteriaAge:Bacteria 3,500,000,000 years (2000x)Eukaryotes 1,800,000,000 years (1000x)Multicellulars 580,000,000 years (300x)Australopithecus 4,000,000 years (2x)Homo erectus <2,000,000 years (1)Antibiotic use 60 years (0.00003x)

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” !!!

They are ubiquitous and indestructibleGould SJ. Life’s Grandeur; Vintage (Pub), 1997.