Surveillance for Antimicrobial Resistance and Preparation of an Enhanced Antibiogram at the Local Level janet hindler
At the conclusion of this talk, you will be able to Describe CLSI M39-A3 recommendations for preparing cumulative antibiograms. Discuss some problems in calculating %S statistics and how to address them. List points to consider when evaluating cumulative %S data.
CLSI M39-A3 Guideline Analysis and Presentation of Cumulative Antimicrobial Susceptibility Test Data M39-A3 2009
CLSI M39-A3 Guideline describes preparation of a cumulative antibiogram report to be used to support clinical decisions re: empiric therapy of initial infections
Cumulative antibiogram Generally one big report, but increasing emphasis on segregating data to answer specific questions
What does it mean to be compliant with CLSI M39-A3? (1) Analyze/present data at least annually Include only final, verified results Include only species with 30 isolates Include diagnostic (not surveillance) isolates Include the 1 st isolate/patient, irrespective of body site overall antimicrobial susceptibility profile Include only drugs routinely tested Calculate %S (do not include %I)
What does it mean to be compliant with CLSI M39-A3? (2) Staphylococcus aureus list %S for all and MRSA subset Streptococcus pneumoniae list %S for cefotaxime / ceftriaxone / penicillin w/ meningitis and non-meningitis breakpoints list %S for penicillin w/ oral breakpoints, if appropriate Viridans streptococcus - list %S and list %I for penicillin
Special Circumstance (1) % Susceptible N Amx Cftrx Ctax Ery Lvx Pen (IV) Pen (oral) Sxt S. pneumoniae 96 96-1 - 1 74 99-1 71 4 74 meningitis* 96-91 2 90 2 - - 71 2 - - nonmeningitis** 96-97 3 97 3 - - 95 3 - - %S calculated using: * meningitis breakpoints **non-meningitis breakpoints cont..
Special Circumstance.. S. pneumoniae %S (2) 1 Breakpoints differ for ceftriaxone, cefotaxime and penicillin based on diagnosis. 2 Susceptible breakpoint for S. pneumoniae in patients with meningitis is 0.5 µg/ml for ceftriaxone and cefotaxime and 0.06 µg/ml for penicillin. 3 Susceptible breakpoint for S. pneumoniae in patients with nonmeningitis infections is 1 µg/ml for ceftriaxone and cefotaxime and 2 µg/ml for penicillin. 4 Susceptible breakpoint for S. pneumoniae is 0.06 µg/ml for penicillin when penicillin is administered by the oral route.
Special Circumstance.. % Susceptible Option 1 N Clin Ery Ox Pen T-S T S Van All SA 1317* 80 50 68 13 97 100 MRSA 449 44 4 0 0 94 100 MSSA 904 97 72 100 18 97 100 * All NOT sum of MRSA and MSSA because: analyzed 1 st /iso/patient from each subset, e.g., OX-R S. aureus 36 pts had both MRSA and MSSA CLSI M39-A3
Special Circumstance.. Option 2 % Susceptible N Clin Ery Ox* Pen T-S T S Van MRSA 449 44 4 0 0 94 100 MSSA 904 97 72 100 18 97 100 *68% of all S. aureus are oxacillin-s CLSI M39-A3
Exmp. combine 2 years of data when N is <30 isolates Morganella morgannii Year N %S Cipro # S %S Gent # S 2006 18 72 13 89 16 2007 22 68 15 86 19 Total N 40-28 - 35 %S = Total #S / Total N %S - 70 (28/40) - 88 (35/40) -
If you must report N <30, consider comment such as %S calculated from fewer than the standard recommendation of 30 isolates
Cumulative antibiogram data are impacted by Patient population served Culturing practices Antimicrobial susceptibility testing policies Temporal outbreaks Assumes individual susceptibility test results used for cumulative antibiogram are Final Verified Results!
Example: Healthcare facilities showing differences in %S for oxacillin and S. aureus %S oxacillin Facility type Geriatrics? Known MRSA outbreaks? Perform AST on all S. aureus? Facility #1 35% Tertiary care Yes Yes No Facility #2 50% Acute care No No Yes if comparing %S data, you must understand the basis of the data!
To interpret cumulative antibiogram, data you should know (1) Source of isolates Random? Special collection; any selection criteria? Patient demographics? Geographic region? Dates of collection? Fresh isolates? Frozen isolates? Could isolates be clonal?
To interpret cumulative data antibiogram, you should know (2) Laboratory testing Method used for organism identification and AST? Criteria used to define susceptible? All drugs tested on all isolates concurrently? Isolates tested by central lab or in individual labs? QC documented? Unusual results confirmed? Data analysis Eliminate duplicates? Does %S include %I? N (small N value has large 95% confidence interval)
What are the options for excluding duplicate patient isolates? Count one isolate of a given species / patient / year M39-A3 suggests 1 st isolate/patient Count duplicates after x days (e.g. 7, 30 days) elapsed since testing the previous isolate from a given patient Count duplicate isolates on a patient that have different results for one or more drugs (e.g., S to R, R to S) Count all isolates
How do various methods for excluding duplicates impact %S? P. aeruginosa - ciprofloxacin # isolates % Cipro-S 2.2 isolates/patient 1600 1200 68 65 67 1571 70 65 800 400 725 862 773 54 60 55 0 1st iso/patient 30 days Phenotype All M39-A3 50 UCLA 2008
When might we examine more isolates than the 1 st isolate/patient? When data from additional isolates are needed to answer a specific question Examples: How many patients had MDR Acinetobacter baumannii? Were any S. aureus not susceptible to daptomycin?
What about MDR Acinetobacter baumannii? A. baumannii - meropenem 1.9 isolates/patient # isolates % Mero-S 200 58 60 150 185 41 50 40 100 50 96 30 20 10 0 1st iso/patient All 0 M39-A3 UCLA 2008
How many Acinetobacter baumannii were MDR? Category # Isolates # Patients Comments All 185 96 # MDR???? 1 st st isolate/pt 96 96 # MDR???? MDR only 110 43 For 3/43 pts, MDR was not 1 st iso/pt MDR = R to: aminoglycosides; carbapenems; anti-pseudomonal cephalosporins, penicillins, fluoroquinolones UCLA 2008
What are some concerns for calculating %S? Not all drugs may be tested on each isolate of a given species When this occurs, comparing results for one drug against another may be misleading Selective testing of isolates from specific body sites
Concern: 1) separate panel for testing Enterobacteriaceae from urine; 2) testing large numbers of urine isolates can dilute data N Am Cz E. coli % Susceptible Cip Lvx Sxt Non-urine only 292 44 82 80 80 62 Urine only 3417 63 93 93 NT 77 All 3636 61 92 92 80 76 Both cipro and levo tested on gram-neg (non-urine) panel; cipro tested only on urine panel CLSI M39-A3
Optional Report Strategy (include comment) N Am Cz % Susceptible Cip Lvx Sxt E. coli 3636 61 92 92 80* 76 Levofloxacin tested on non-urine isolates only (N=292). Levofloxacin results should not be compared to results of other drugs, all of which were tested on all isolates (urine and non-urine, N=3636). CLSI M39-A3
Concern: reporting % VRE?? Protocols for AST of enterococci may vary among labs (UCLA tests sterile site isolates, urine isolates from inpatients, and on request) N Van % Susc Enterococcus spp. - bld 143 51 Enterococcus spp. - all 787 70 UCLA 2008
What about stratifying %S data to provide suggested empiric therapy recommendations for select patients / conditions? By specimen type or infection site By nursing unit or site of care By organism s s resistance characteristics By clinical service or patient population
Example of %S Data Stratification using cumulative antibiogram data to guide empiric therapy of acute uncomplicated urinary tract infections (UTIs)
Organisms Tested for AST UCLA 2008 (N = 15,645) E. coli NOT UR E. coli UR P. aeruginosa K. pneumoniae Other Other GP Other GN Enterococcus GN, gram-negative bacteria GP, gram-positive bacteria S. aureus
Managing Uncomplicated UTIs in Women Risk factors for resistance? Current or recent TMP-SMX or other antibiotic Recent hospitalization Recurrent UTI in past year No TMP-SMX resistance in community >20%? Yes Yes Use alternative agent: Fluoroquinolone (3 days) Or Nitrofurantoin (7 days) Fosfomycin (single dose) No TMP-SMX (3 days) Warren et al. 1999. CID. 29:745. Warren et al. 1999. CID. 29:745. Gupta, Hooton, and Stamm. 2001. Ann Intern Med. 135:41.
Variability in Urine C&S Ordering Practices Clinical Scenario Probable uncomplicated UTI Previous treatment failure in older woman 1 No. (%) of general practitioners who said they would order C&S 165/278 (59%) 262/291 (90%) 1 more likely to have resistant organisms Hillier et al. 2006. J Antimicrob Chemother. 58:1303. Hillier et al. 2006. J Antimicrob Chemother. 58:1303. Patient with uncomplicated UTIs often not cultured!
E. coli Urine Isolates % S data from 2 facilities during similar time frame Data Source 10/11/04 1/31/05 CA university health care service 1 All pts. presenting with uuti symptoms Outpatients 3 116 3308 Cip 97 2004 UCLA Cumulative Antibiogram Report 2 2 %S from first isolate/pt (CLSI M39-A3) 3 Includes ER patients N 88 TMP-SMX 84 76 1 Smith et al. 2008. Clin Infect Dis. 46:689.
E. coli Urine Isolates %S 1 4 Patient Populations Category All patients Inpatients Outpatients 2 18-40 yo female outpatients 2 N 3801 421 3411 998 Am 53 40 54 60 Cz 88 75 89 92 Cip 79 62 80 92 Fm 99 98 99 99 T-S 72 61 73 77 1 %S from first isolate/pt (CLSI M39-A3) 2 Includes ER patients UCLA 2008
E. coli Urine Isolates (N=998) 18-40 yo Female Outpatients Resistance Profile N % No Resistance Am Am Sxt Am Cz Am Cip Sxt Am Cz Sxt Sxt Am Cip Am Cz Cip Sxt 561 139 127 38 34 27 18 12 11 56.2 13.9 12.7 3.9 3.4 2.7 1.8 1.2 1.1 Drugs examined: Am, Cz, Cip,, Fm, Sxt Cip Other 9 22 0.9 2.2 UCLA 2008
Example: empiric therapy policy change following clinical failure and review of cumulative antibiogram data Elderly male patient with cystitis seen in ER failed empiric ciprofloxacin Rx; previously empiric Rx successful Performed culture (previously no culture) which revealed cipro-r E. coli ER MD Question: what is cipro %S for urine isolates of E. coli from ER patients?
E. coli Urine Isolates %S ER vs. All Outpatients N Am Cz Cip Fm Sxt ER <65 yo 329 52 89 85 100 64 ER 65 yo 140 45 89 62 96 64 All 3103 59 92 84 99 74 Revised ER Policy; use cipro empirically only in patients <65 yo UCLA 2005
Stratifying %S Data for MDR Organisms If MDR (e.g., KPC) endemic, consider stratifying data (e.g., by unit or R mechanism) Minimize skewing of data Example: KPCs might be predominantly among ICU patients May not be seen in OB-GYN; single cumulative antibiogram would offer limited guidance for empiric therapy options for OB-GYN patients
Stratifying %S Data for MDR Organisms (e.g., ESBL, KPC) CLSI M39-A3
What about %S data for combinations of agents? Guide empiric therapy of infections where the likely causative organisms are best treated with combinations of antimicrobial agents Calculate %S to either (or both) of the two agents Does NOT imply synergy, antagonism or likely activity in vivo
P. aeruginosa % Susceptible* (N=726) Option 1 Cip Cftaz Mero Tob Cftaz +/or Cip Mero +/or Cip Cftaz +/or Tob Mero +/or Tob 68 81 83 87 89 87 94 93 94% S includes isolates that have any of these 3 profiles Cftaz S S R Tob S R S *analysis included 1 st isolate/patient UCLA 2008
P. aeruginosa (N=726 patients 1 ) Option 2 %S to 7 drugs and %S to either or both when 2 drugs are evaluated (%S for 2 drugs does NOT imply synergy, antagonism or likely activity in vivo) Ceftaz (74%) Mero (79%) Pip-taz 4 (69%) Cip (64%) Amik (94%) 2 98% 3 98 98 97 Gent (80%) 93 93 91 88 Tob (85%) 94 93 92 89 Cip (64%) 1analysis included the most R result for each drug if patient had >1 isolate 2 %S for one drug - value in ( ) 3 %S for either or both drugs (e.g., %S to amik and/or ceftaz) UCLA 2008 4 used 16 µg/ml breakpoint for S 89 87 85 -
Patient JWR had 4 isolates of P. aeruginosa Isolate # 1 2 3 4 Cip S R R R Gm R R R R Mero S S S S Most R result for each drug R R S
UCLA Resistance Trends (from UCLA Antimicrobial Susceptibility Summary) 60 50 % Resistant 40 30 20 10 0 1990 1994 1998 2000 2002 2003 2004 2005 2006 2007 2008 MRSA PSA-Cip VRE E. coli-cip UCLA
Cumulative Antibiogram QA Check Ensure individual AST results on pt. isolates are verified before being sent to database Report species with 30 isolates Define abbreviations Include appropriate drugs for the species Only those with breakpoints in CLSI M100 for the species Check footnotes (e.g., only ampicillin, a fluoroquinolone, TMP-SMZ, 3 rd gen cephalosporin, chloramphenicol for Salmonella / Shigella)
Cumulative Antibiogram QA Check (con t) Get 100% S for drug/bug combinations with only S breakpoints in CLSI M100 (e.g., Streptococcus pneumoniae and vancomycin) Staphylococcus - %S OX should = %S for other β-lactams (except amp, amox,, pen which should have a lower %S) Ensure results for related drugs make sense Compare data to previous year; account for significant changes Compare local data to national data
Trouble-Shooting N % Susceptible Imip Vanco K. pneumoniae 230 85 - S. aureus 781-98
Trouble-Shooting N % Susceptible Imip Vanco K. pneumoniae 230 85 a - S. aureus 781-98 b a Were there outbreaks with R strain (e.g., KPC producer)? b Are the 2% (16 isolates) that were VISA or VRSA accounted for?
Tools for examining statistical significance of %S Table H1. 95% confidence intervals for selected sample sizes CLSI M39-A3
How reliable is a report of 80% S for E. coli and ciprofloxacin? 95% Confidence Intervals for %S with Selected Sample Sizes Sample size 10 100 1000 80% S 44-99 71-87 77-82 CLSI M39-A3
Tools for examining statistical significance of %S Table H2. For comparing changes in %S CLSI M39-A3
If 80% of E. coli are susceptible to ciprofloxacin, what decrease in %S would be statistically significant (P 0.05)? # Isolates % S 80% 100 66% 600 75% CLSI M39-A3
Where can we find %S data from around the globe? Publications CDC / state Public Health Dept websites Pharmaceutical company websites EARRS Other Some examples follow!
http://www.cdc.gov/narms/index.htm
http://www.health.state.mn.us/divs/idepc/dtopics/antibioticresistance/antibio2007.pdf tance/antibio2007.pdf
http://www.gpcsusceptibilitydata.com/content/home.asp?page www.gpcsusceptibilitydata.com/content/home.asp?page=home
http://www.wyeth.com/hcp/tygacil/test
http://www.escmid.org/research_projects/eucast/documents/other_documents/ ocuments/
http://217.70.33.99/eucast2/searchcontroller/regshow.jsp?id=1212
http://www.rivm.nl/earss www.rivm.nl/earss/
Note: search for publications only http://www.jmilabs.com/default.cfm
http://www.doctorfungus.org/index.htm
Summary (1) CLSI M39-A3 provides guidelines for preparation of a cumulative antibiogram to guide physicians in empiric therapy of initial infections. Cumulative antibiogram data are impacted by several factors to include: 1) culturing practices; 2) susceptibility testing policies; and 3) patient population served. The method used to exclude duplicate isolates on a given patient can impact cumulative antibiogram data. If drugs included in the cumulative antibiogram are only tested on select isolates,, data may be skewed.
Summary (2) It may be helpful to stratify %S data by one or more variables when developing empiric therapy guidelines for select patient populations, infections, etc. %S data for two drugs (e.g., %S to either of the two or both) may be helpful for guiding empiric therapy for suspected or confirmed infections caused by organisms generally treated with multiple agents. If the number of isolates is small,, the 95% confidence interval will be large. CLSI M39-A3 contains tools to help determine the statistical significance of changes in %S between time frames.
Summary (3) Each facility must customize cumulative antibiogram reports to optimize their use in that facility. There are several national and international sources of %S data. It is important to know source/nature of data used to construct the database when drawing conclusions. Although there are written guidelines for analysis and presentation of cumulative antimicrobial susceptibility test data, guidelines for use of these data are limited.