Session One:

1
Session One Introduction to Good Antimicrobial
prescribing Antibiotic Selection Laboratory
based perspective Alasdair MacGowan Bristol
Centre for Antimicrobial Research Evaluation
(BCARE) University of Bristol/North Bristol NHS
Trust Southmead Hospital
2

• What can laboratory data inform?
• For the individual patient
• identification of the potential pathogen
  antibiogram
• site of infection for definitive therapy
• part of a risk assessment for empiric therapy
  i.e. carriage
• of a resistant isolate
• therapeutic drug monitoring, reduce toxicity
  risk optimise
• microbiological outcomes
• Epidemiological information
• risk factors for resistance evolution, extent,
  geography of resistance
• local, national, international
• based on lab generated data based on lab
  derived clinical isolates

3

• Key issues for antibiotic selection are
• data is epidemiologically relevant
• predict clinical outcomes

4
Epidemiologically relevant data? Collection of
resistance data in the community
patients with infection patients presenting with
infection nurse specimen sent to
laboratory laboratory data (location, specimen
type, potential pathogen antibiogram)
other (NHS Direct) (Casualty) (? OTC)
GP
5

• The pitfalls
• data skewed because patients who fail initial
  therapy
• more likely to be tested
• age bias towards young and old
• isolates from different diseases included
  disproportionately,
• i.e. sinusitis vs LRTI, exacerbation COPD vs
  CAP
• variation in methodology (CLSI vs EUCAST
  breakpoints)

Livermore, MacGowan, Wale, 1998
6

• Examples of bias (1)
• 1 in 35 episodes of chest infection results in
  a sputum referral
• sputum referral not age matched with LRTI
  diagnosis

Lovering, unpublished
7

• Examples of bias (2)
• not all specimens sent relate to infection
• lt50 MSU relate to UTI
• Impact on resistance assessment
• lab databases over estimate resistance
• ampicillin resistance in H.influenza 20 in lab
  database
• 11 in unselected patients who receive
  antibiotics
• MacGowan et al, 1998
• trimethoprim resistance 12-17 in E.coli from
  patients with frequency
• dysuria syndromes 25 in lab database
• McNulty, Lovering, unpublished

8

• Hospital based surveillance
• generally on firmer ground i.e. bacteraemia
  data
• National resistance in bacteraemic isolates

9
Informs collection of local data
10
Informs collection of local data
11
Does susceptibility testing predict outcome?
12
Adequate and inadequate antimicrobial
chemotherapy
adequate therapy which is effective against the
micro-organism causing infection (based on
susceptibility testing) inadequate microbiologica
lly documented infection which is not being
adequately treated as? drug is not active
against the pathogen? pathogen is resistant to
the drug (based on susceptibility testing)
13
Location ICU studies (1)
14
ICU studies (2)
15
Resistances Extended spectrum Blactamase
producing E.coli/Klebsiella (1) Kim et al,
2002142 bloodstream isolates in Korea, E.coli or
K.pneumoniae, strains MIC gt2mg/L to third
generation cephalosporins Patients treated with
extended spectrum cephalosporin
16
Does susceptibility testing predict
outcome? YES NO

• ICU
• intra abdominal sepsis
• MRSA
• ESBL production
• P.aeruginosa

• penicillin non susceptible pneumococci
• P.aeruginosa

17
Understanding the discrepancies
wild type distribution
wild type cut off
susceptibility result (MIC, mg/L)
S.pneumoniae
Pharmacodynamic index (TgtMIC Cmax/MIC AUC/MIC)
MRSA ESBL resistances which ?? MIC
Pharmacokinetics
P.aeruginosa efflux pumps VISA
Microbiological outcome
Clinical breakpoint
Clinical outcome
18

• Laboratory based perspective on antibiotic
  selection
• provides epidemiological data for empirical
  therapy
• (beware bias)
• provides individual patient data to optimise
  outcome
• (needs to be placed in the correct context)
• provides strains for further study