International Interlaboratory Proficiency Testing Program for Measurement of Azole Antifungal Plasma Concentrations: a Five-Years Data Analysis V.J.C. Lempers, J.W.C. Alffenaar, D.J. Touw, D.M. Burger, D.R.A. Uges, R.E. Aarnoutse, R.J.M. Brüggemann 1st International Workshop on Clinical Pharmacology of Antifungal Drugs and Fungal Diseases Berlin, 26 th of April 2013
International Interlaboratory Proficiency Testing Program Wide application of analytical methods to measure azole antifungal plasma concentrations Validation tool when analysing antifungal plasma concentrations and enhance methods accuracy, precision and specificity. Proficiency testing program started in 2007: first initiative of an international interlaboratory QC program for the measurement of azole antifungal agents Rounds organized by the Dutch Association for Quality Assessment in Therapeutic Drug Monitoring and Clinical Toxicology (KKGT).
Methods overview (1 round) Establish High/Low conc. Spike Fluconazole Itraconazole OH-itraconazole Voriconazole Posaconazole Flucytosine (since 2010) Obtain plasma Spiked Human Plasma Perform confirmative check with own HPLC (<5% target conc.) Send high&low spiked plasma to participating lab Repeat round (2/year) Report results anonymously & give feedback on performance Calculate deviation (-20% to +20% considered to be correct) Analysis (< 6 weeks) + details on analysis
High/Low concentrations per year 35 30 FLZ and VCZ concentrations (high/low) per round FLZ VCZ Concentrations (mg/l) 25 20 15 10 5 0 Round (year/round/high-low)
Worldwide participation of laboratories Netherlands USA United Kingdom Switzerland Spain Italy Belgium Germany France Canada Sweden Korea, South Australia 2008 2012 2008 2012 Europe 34 44 North America - 11 Asia - 1 Australia 1 1 Total 35 57 (+38.6%) 0 5 10 15 20 Number of laboratories
HPLC/LC-MS dominate methods Percentage 60 55 50 45 40 35 30 25 20 15 10 5 0 55.0 43.4 1.6 HPLC LC-MS Other Method Absolute Percentage HPLC 1238 55,0 LC-MS 977 43,4 GC-MS 5 0,22 UPLC 31 1,38 Total 2251 100%
HPLC / LC-MS analyses per drug 450 HPLC/LC-MS per drug (absolute) Number of Analyses 400 350 300 250 200 150 100 50 0 FLZ ITZ OH-ITZ PCZ VCZ FLUC Antifungal HPLC LC-MS HPLC overall preferred method of analysis (except for posaconazole) Flucytosine: Almost no labs use LC-MS (5.5%) Voriconazole analyzed the most (29%), flucytosine the least (4.9%)
Analysis per drug/year Analyses 200 150 100 50 Analyses/drug per year FLZ ITR OH-ITZ PCZ VCZ 0 2008 2009 2010 2011 2012 Year Increase in number of analyses per year for each drug More participations over time Development of new analytical methods FLUC
1 out of five analyses outside 80-120% limit Percentage 100 90 80 70 60 50 40 30 20 10 86.94 Analyses within 80-120% 2008-2012 76.08 75.30 71.96 85.45 74.25 0 FLZ ITZ OH-ITZ PCZ VCZ FLUC Azole antifungal 80,8% of analyses within 80-120% (7,7% >120% and 11,5% <80%) (Very weak) negative correlation between relative inaccuracy and year of the program (r = -0.054, n=2251, p=0.012).
Hydroxy-itraconazole is most difficult to determine adequately Antifungal Mean ab. Inacc. (%) N FLZ 12.9 287 ITZ 29.6 451 OH-ITZ 34.4 348 PCZ 30.0 402 VCZ 12.1 652 FLUC 13.0 111 Significant effect of antifungal drug on absolute inaccuracy was found Laboratory, method of analysis or concentration did not influence results [4-way ANOVA, F=3.58; P=0.003] LOW concentration (ITZ, hitz, PCZ) yield a worse than average performance. HIGH concentrations (FLZ, FLUC), have a better performance.
HPLC more frequently out of bounds (absolute) Percentage 25% 20% 15% 10% 5% 0% Analyses outside 80-120% (absolute) 21,3% HPLC 15,8% LC-MS Method Mean ab. Inacc.(%) N HPLC 22.3 1238 LC-MS 19.3 977 GC-MS 16.4 5 UPLC 18.4 31 LC-MS significantly better in total analyses within 80-120% (χ2 (1, N = 2215) = 11.035, p <.005). Mean absolute inaccuracy of 22.3% for HPLC and 19.3% for LC-MS NOT significantly different (n = 2215, p=0.379, unpaired t-test)
High concentrations can be determined more accurately 35 Percentage outside 80-120% 30 Percentage 25 20 15 10 Low High 5 0 FLZ ITZ OH-ITZ PCZ VCZ FLUC Analyses at low concentrations significantly less accurate than for the higher concentrations (75.5% versus 86.6% correct analyses, respectively; p<0.001, fourway ANOVA). Methods used have LLOQ higher than low concentration for specific antifungal?
Results error forms Explanation for inaccuracies 0 5 10 15 20 M5 Imprecision due to result being close to detection limit T16 Concentration below quantitation limit C1 Results not correctly transcribed to report form C2 Result reported for the wrong sample M9 Method is biased M17 Other method problem (please specify in the comment T24 Other technical problem (please specify in comment box) Unexplained/unassigned cause T6 Material improperly prepared/stored T17 Inadequate equipment maintenance M11 Method lacks specificity T14 Aging stock solutions M14 Method used without validation C12 Other (please specify in comment box) M4 Problem in manufacture of reagents or reference materials # of explanations Error forms based on CLSI guidelines: 23 respondants Sources of error: Clerical (C), Methodological (M), Technological (T), Equipment (E), Organizational (O) or other
Conclusions After 5 years of analysis by other laboratories: From 33 to 57 laboratories, still growing HPLC and LC-MS mostly used as method of detection (55.0 and 43.4%, respectively) Increase in number of analyses/year within 80-120% margin over 5 years LC-MS significantly more analyses within 80-120% High concentrations measured significantly more accurate Explanation for inaccuracies: concentration below LLOQ/ close to LOD
Future prospects Program always open for new participants (costs 250/year, 2-rounds) Recruit more laboratories Include more antifungal drugs (alphabetical order) Echinocandines Anidulafungin Caspofungin Micafungin Newer azoles?
Acknowledgements Radboud University Nijmegen Medical Centre, the Netherlands Dr. R.J.M. Brüggemann Prof. Dr. D.M. Burger Dr. R.E. Aarnoutse University Medical Centre Groningen, the Netherlands Dr. J.W.C. Alffenaar, Prof. Dr. D.R.A. Uges Association for Quality Assessment in Therapeutic Drug Monitoring and Clinical Toxicology (KKGT), The Hague, the Netherlands Dr. D.J. Touw Analytical staff Association for Quality Assessment in Therapeutic Drug Monitoring and Clinical Toxicology & UMC St. Radboud