Quality Assurance of Antimicrobial Susceptibility Testing Gunnar Kahlmeter

Quality Assurance of Antimicrobial Susceptibility Testing Gunnar Kahlmeter Dept of Clinical microbiology, Växjö, Sweden The National Institute for Infectious Disease Control SRGA & SRGA-M, BSAC and EUCAST

Antimicrobial susceptibility testing is about short-term and long-term credibility! Surprising findings or trends will be questioned!

Standardisation/Calibration of method Education Quality Assurance Internal Quality Assessment (specimen reprocessing) Routine quality control Expert rules External Quality Assessment Adapted from Derek Brown

Standardisation and calibration of method • Standardised methodology (BSAC, CASFM, DIN, NCCLS, SRGA) • QC-strains with target values • Reference MIC-distributions of wild type bacteria (on www. eucast. org) • Reference zone diameter distributions of wild type bacteria (on SRGA website and within the BSAC system)

E. coli ATCC 25922 Antimicrobial Uppdaterad 2000 -04 -18 MIC (mg/l) Disc (µg) ISA (mm) Ampicillin Mecillinam Piperacillin 2 -8 0. 06 -0. 25 1 -4 10 10 30 17 -23 31 -37 24 -30 Pip/tazobactam Cefadroxil Ceftibuten Cefuroxim Cefotaxim Ceftazidim Aztreonam Imipenem Meropenem 1/4 -4/4 4 -16 0. 12 -0. 5 2 -8 0. 03 -0. 12 -0. 5 0. 06 -0. 25 0. 008 -0. 03 30/6 30 30 30 10 10 24 -30 16 -22 33 -39 23 -29 37 -43 33 -39 36 -42 32 -38 35 -41

Reference MIC-distributions on www. eucast. org

Reference MIC-distributions on www. eucast. org

Reference zone diameter distributions through SRGA and BSAC Data by Trevor Winstanley

Reference zone diameter distributions through SRGA and BSAC

E. coli vs. gentamicin 30 µg n=2478 from 17 countries S/R 21/17

Discrepancies discovered in the validation procedure - analysis and suggested measures - one or more of the distributions deviate from what is expected (the median is off, the distribution is wider than expected, the part of the distribution that ought to be unimodal is bimodal, etc) the pattern of the deviation can be used to diagnose the problem: • are all basic parameters alike for the two histograms (disk strengths, additives, atmosphere, calibration of the CO 2 -incubator, temperature etc). • are species-identification procedures equivalent? • do the aberrations occur with media containing supplements? or with media or disks sensitive to storage? or with antibiotics sensitive to changes in p. H (aminoglycoside, erythromycin)? or with antibiotics sensitive to the autoclaving of the medium (nitrofurantoin).

Zone diameter distributions from 4 Swedish laboratories illustrating different methodological ”outputs”. Reference SRGA distribution Median

Discrepancies discovered in the validation procedure - analysis and suggested measures If the majority of the distributions deviate from the expected there should be a pattern: • inhibition zones smaller than the reference histograms: inoculum too dense or incorrect agar volume in the plates (depth >4 mm). • inhibition zones larger than the reference histograms: incorrect agar volume in the plates (depth <4 mm).

Standardisation/Calibration of method Education Routine quality control Quality Assurance Internal Quality Assessment (specimen reprocessing) Expert rules External Quality Assessment Adapted from Derek Brown

Routine quality control 1. QC-strains with target values (target values, permitted range, trend analysis in Shewart diagrams) - general: overall proficiency check - specific: rare organisms N. gonorrhoeae, H. pylorii, fungi etc 2. Serial analysis of zone diameter distributions (or MIC-distributions)

E. coli ATCC 25922 Antibiotikum Uppdaterad 2000 -04 -18 MIC (mg/l) Disc (µg) ISA (mm) Ampicillin Mecillinam Piperacillin 2 -8 0. 06 -0. 25 1 -4 10 10 30 17 -23 31 -37 24 -30 Pip/tazobactam Cefadroxil Ceftibuten Cefuroxim Cefotaxim Ceftazidim Aztreonam Imipenem Meropenem 1/4 -4/4 4 -16 0. 12 -0. 5 2 -8 0. 03 -0. 12 -0. 5 0. 06 -0. 25 0. 008 -0. 03 30/6 30 30 30 10 10 24 -30 16 -22 33 -39 23 -29 37 -43 33 -39 36 -42 32 -38 35 -41

Period of calibration

Variation attributable to the lab. technician (the inoculum) One of my oldest friends – but her inocula were always a bit too dense!

2/25/2021 G Kahlmeter, SRGA & SRGA-M

Discrepancies in results with quality control strains - analysis and remedies Random errors cannot be avoided but shall be controlled. Reference strains have been assigned target values and acceptable intervals for random variation. Occasional deviation can be accepted but trends and recurring patterns should be investigated, explained and corrected.

Discrepancies in results with quality control strains - analysis and remedies Systematic errors (systematic high or low values, or a trend) should be investigated and corrected. - wrong strain - disc content - volume of medium - p. H of medium - atmosphere and incubation temperature) - aging of plates If this does not explain and solve the problem the most probable cause is the density of the inoculum.

A change in the ion-content of ISA (Oxoid)

Routine quality control 1. QC-strains with target values (target values, permitted range, trend analysis in Shewart diagrams) - general: overall proficiency check - specific: rare organisms N. gonorrhoeae, H. pylorii, fungi etc 2. Serial analysis of zone diameter distributions (or MIC-distributions)

Internal QA Serial analysis of zone diameter distributions The regular analysis of zone diameter distributions of consecutive clinical isolates provide good internal quality asessment: median (accuracy) width (reproducibility) shape (tells you all sorts of things about the drug, species, lab)

A high degree of reproducibility! A high degree of credibility!

Epidemic in the Lagan area

Standardisation/Calibration of method Education Routine quality control Quality Assurance Internal Quality Assessment (specimen reprocessing) Expert rules External Quality Assessment Adapted from Derek Brown

Expert rules are an important part of QA Penicillin resistance does not exist in S. pyogenes - ”faulty disk” Ampicillin resistance in E. faecalis is very rare - ”erroneous species identification” Methicillinresistance in Staphylococci should not permit an ”S” for other betalactam antibiotics - ”don´t do them – you´ll get it wrong!” Mistrust clarithromycin ”S” if erythromycin ”R” - ”someone got the breakpoint wrong”

Standardisation/Calibration of method Education Routine quality control Quality Assurance Internal Quality Assessment (specimen reprocessing) Expert rules External Quality Assessment

External Quality Assessment • The distribution of strains with known antimicrobial resistancies or MIC-values - NEQAS, EQUALIS, RING, LABKA etc - panel of experts to chose strains and to make the analysis of results and prepare comments • The analysis of yearly distributions of the quantitative results of consecutive clinical isolates – combined EQA and resistance surveillance.

EQA in NEQAS and EARSS - please refer to NEQAS and EARSS reports

Sweden • 100 consecutive patient isolates per bacterium and antibiotic and year and lab. • Each lab enter zone diameters over internet. • QA - the distributions are compared (median, width, scew) with reference. • Resistance surveillance - using SRGA breakpoints the resistance frequencies calculated and tabulated

S/R 18/17 • Each laboratory is asked to test 100 consecutive patient strains of defined species (3 - 5) against defined antibiotics (4 - 6). • The zones are entered over the internet (see www. srga. org) • 30 labs x 100 strains = 3000 strains per species & antibiotic & year with good geographical distribution.

Res. Net