Validation Manager* version: 2.18I ©VWR INTERNATIONAL SAS, France software and Microsoft. Excel. Linearity of this method was confirmed by linearity tests ...
I
CONFERENCE
© RAVIMED S p. z o. o. 2008
OF
POLISH MASS SPECT ROMETRY SOCIET Y 16-18
IV
2008
Puławy
NEW SENSITIVE LCLC-MS METHOD FOR DIGOXINE DETERMINATION IN HUMAN PLASMA VALIDATION AND QUALITY ASSURANCE DURING STUDY Tomasz Grabowski, Anna Świerczewska, Beata Borucka, Renata Sawicka, Wiesław Raszewski Centrum Badań Farmakokinetycznych FILAB, Ravimed Sp. z o.o. Łajski www.filab.com.pl
INTRODUCTION
GOAL
Digoxin is a drug used in cardiac diseases. The range of therapeutic concentration for digoxin is estimated from 0,8 to 1,6 ng/ml, but already 2 ng/ml is consider as toxic. In that reason very sensitive and accurate method of determination is needed. In the past the immunochemical methods were applied to determine digoxin concentration. Those methods however are described as not selective and not accurate enough. FDA not recommends immunochemical methods for pharmacokinetics study. For this specific drug we also have to deal with DLIF element. Therefore the HPLC selective and sensitive method with MS/MS detection for determination of digoxin in human plasma was developed.
The aim of this study was to develop method for determination of digoxine in human plasma in addition to internal standard (I.S.) area – digitoxin which would be selective, sensitive and robust enough to performed high number of analyses.
RESULTS For this bioanalytical method full validation were performed. The following validation parameters were estimated: linearity, precision (repeatability and intermediate precision), accuracy, recovery (extraction efficiency), limit of the quantitation (LLOQ), selectivity and carry over. The best calibration curve fitting achieved for the linear regression y=a+bx. All calculations were process by Validation Manager* version: 2.18I ©VWR INTERNATIONAL SAS, France software and Microsoft Excel. Linearity of this method was confirmed by linearity tests (F1 and Mandel) and correlation coefficient (R2) which for three calibration curves was higher, for all of them, the 0,995 and for curves from the study mean R2 was 0,998 with RSD 0,1%. Also the slope for all calibration curves (n=16) hasn’t shown significant differences between them (mean slop 35,18 with RSD 5,66%). For all calibration points the repeatability and intermediate precision presented as RSD was below 10% and the means equal 5,04% and 5,97% respectively. Accuracy was established over the range of linearity base on the data from linearity study. Presented as the mean recovery of theoretical concentration equals 100,2% with RSD 2,61%. Recovery was presented as efficiency of extraction comparing data from water solution and extracted samples far all calibration points (n=27; 9 calibration points). The mean efficiency of extraction equals 102,0% with RSD 6,42%.The lowest limit of quantitation (LLOQ) was determined from linearity experiments and was set to 0,105 ng/ml with RSD 7,94% (n=9). Also the lowest limit of detection was determine based on experiment data and was set to 0,025 ng/ml. Selectivity was confirmed by sowing no significant peaks at the retention time of digoxin and I.S. in six blank plasma samples. The chromatographic system was tested, in order to evaluate the possibility of carry-over, by alternatingly injection of the highest limit of quantitation (HLOQ) and the blank plasma. The contribution of HLOQ to the response of the blank plasma was below 0,1% of the response of the HLOQ. The stability tests were performed on quality control concentration. The autosampler stability of extracted samples stored at 10°C was estimated for 48h, the working standard solution stability stored at 2-8°C for 14 days and long term stability of samples stored in low temperature freezer at -75°C for 130 days. Tolerance of analytical samples for thawing and freezing cycles was also verified. Stability tests confirmed that no deterioration occurred during two freezing and thawing cycles. For all stability results the deviations from start data were less then 15%. To guarantee the proper quality of obtained results the System Suitability Test (SST) was performed for all chromatograms and the following parameters were fulfilled by them: signal/noise (S/N)>10, peak height/ peak area ratio (H/A)≥5, peak skew (Pk Skew)±1,5. Although analyses were preformed on MS/MS detector all separations were conducted with SIR mode. It was caused by the fact that there was no stabile ion after fragmentation and higher and stable signal were obtained for maternal ions.
METHOD & EQUIPMENT The HPLC System was composed of separation module Alliance 2695 (Waters) with autosampler, degasser and column’s thermostat. The chromatographic separation was performed on Waters Acquity UPLC C18 column (1,7µm 2,1×100mm). The analytes was detected by Waters Micromass Quattro micro MS/MS detector. The data were collected and analyzed by MassLynx 4.1 Waters software. Digoxin and I.S. standards were obtained as pure substance with known assay. All other chemicals reagents were HPLC or S quality. Samples of Li-heparin plasma were obtained either from blood donation center (blank plasma) or from volunteers which participated in clinical part of study (analytical samples). All plasma samples were stored in low temperature freezer at -75°C. A stock solution of digoxin and I.S. were prepared by dissolving and adequately diluting weighed amount of analytical standards in methanol to receive the concentration range from 0,1 to 3,0 ng/ml. Quality control samples were prepared from drug free human plasma in three concentration levels (0,200, 1,50 and 2,50 ng/ml). To 1 ml of plasma sample 30 µl of the I.S. solution (0,500 µg/ml of digitoxin) were added and mixed for 5 s. This mixture was extracted with 3,5 ml of ether by shaking for 30 minutes at 250 rotation per minute and separated by centrifuging for 10 minutes at 4,075×103 g at 20°C. The organic layer was evaporated at 40°C under a stream of nitrogen. Dry residue was reconstituted in 50 µl of re-dissolving solution (acetonitrile). Before injection in the LC system sample was filtered by syringe filter Titan PTFE 0,45 µm. The chromatographic separation of analytes was accomplished by gradient elution with 0,001M ammonium trifluoroacetate and acetonitrile, and flow 0,1 ml/min, during which the percentage of acetonitrile increase from 5% to 90%. With column temperature set on 35°C the retention times for digoxin and I.S. were 5,2 min and 7,2 min, respectively. The settings for ESI ion source were as follow: the capillary temperature was set on 258°C, the capillary voltage 2,90kV, source temperature 80°C and the cone voltage 18,0V. The ESI source was working in negative mode, producing negatively charged ions in the form [CF3COO-] adducts. From ions generated in this way those with ratios m/z 893,5 and 877 were analyzed farther as ions representing digoxin and I.S. respectively. 98
Digoxine
F2:SIR of 4 channels,ES893,5
5.15
2.232e+004 %
6.72
5.96
6.93
7.46
7.66 7.94
5.77
8.40 8.85
9.40
-2
98
Digitoxine
Number of samples in study versus Height/Area ratio of internal standard
F2:SIR of 4 channels,ES-
7.01
877 4.347e+005
%
30
SS
CC
QC
25 5.00
6.00
7.00
8.00
9.00
min
Figure 1 Typical chromatogams of analytes: digoxine – parent dug; digitoxin – internal standard
CONCLUSIONS During digoxin study about 1700 analyses were performed with no significant changes in calibration curves parameters (slope, R2) or in SST parameters. Also after one and the half month break in analyses, reanalyzed samples has shown no significant differences. The slop of new obtained curve comparing to the mean slop from study increase less than 1% (0,3%). It proves that this method is robust, selective and sensitive enough for determine of digoxin concentration as low as 105 pg/ml. This study was performed in accordance to GLP standards and FDA and ICH regulation in certificated laboratory. This method has been successfully applied to first phase of clinical study of the drug.
Height/Area ratio of internal standard
-2
20
15
10
5
0 0
200
400
600
800
1000
Number of samples (SS-study samples; CC - calibation curve; QC - quality conrol)
1200
1400
1600
