| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Drug Monitoring and Toxicology |
1 Laboratoire de Pharmacologie et Toxicologie, Hôpital Maison Blanche, CHU de Reims, France.
2 Centre Régional de Pharmacovigilance, CHU de Reims, France.
3 Laboratoire de Pharmacologie et de Pharmacocinétique, UFR de Pharmacie, Reims, France.
aAddress correspondence to this author at: Laboratoire de Pharmacologie et Toxicologie, Hôpital Maison Blanche, CHU de Reims, 45, rue Cognacq-Jay, 51092 Reims cedex, France. Fax 33-3-2678-8456; e-mail ghoizey{at}chu-reims.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Methods: After standard liquid–liquid extraction with glisoxepide as an internal standard, 8 sulfonylureas (glibenclamide, glipizide, gliclazide, glibornuride, glimepiride, carbutamide, chlorpropamide, and tolbutamide) were eluted from a C18 column within 10 min with an isocratic mobile phase. Drugs were identified and quantified in full-scan MS-MS mode by use of a homemade MS-MS library. We used the assay in 134 cases of hypoglycemic crises of unknown origin.
Results: No ion suppression effect was noted for the analytes at their specific retention-time windows. For all drugs, assay validation showed good linearity (r2 >0.990) and acceptable imprecision and recovery based on commonly used criteria of acceptance. The mean extraction recoveries were 63%–87% for 5 sulfonylureas but <45% for 3 (carbutamide, chlorpropamide, and tolbutamide). Nevertheless, the high sensitivity of the MS instrument made possible detection and quantification of all 8 drugs at subtherapeutic to toxic concentrations with good precision. Sulfonylureas were found in 9 hypoglycemic patients.
Conclusion: The described assay method allows accurate, rapid identification and quantification of 8 sulfonylureas in human plasma and can be used for specific diagnosis of factitious hypoglycemia caused by ingestion of these drugs.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Several analytical methods for the screening and measurement of sulfonylureas in biological fluids have been described. Most methods are based on HPLC with ultraviolet (3), diode array (4), or fluorescence detection after derivatization of serum extracts (5). A micellar electrokinetic capillary chromatographic method with ultraviolet detection has also been proposed for the detection of sulfonylureas in urine (6). These methods may lack specificity, particularly when a single ultraviolet wavelength is used, which may cause false positives. Paroni et al. (7) published an interesting capillary electrophoresis method but clearly mentioned that it should not be used alone to give a definitive diagnosis of intake of these drugs. More recently, higher specificity and sensitivity have been achieved with liquid chromatography (LC)1 techniques with mass spectrometry (MS) used for detection. Magni et al. (8) described such a method that allows the simultaneous identification and quantification of 4 sulfonylureas in serum by electrospray LC-MS. The technique presented by Susanto and Reinauer (9) may lack specificity because only 1 ion per compound for the selected-ion mode detection was used. Recently, Maurer et al. (10) reported a procedure for screening, identification, and quantification of several sulfonylureas in plasma. Analysis was by an atmospheric pressure chemical ionization LC-MS equipped with a single quadrupole. Sulfonylureas were detected by analysis in positive full-scan mode recorded at 2 fragmentor voltages, followed by library-assisted identification using a home-made LC-MS reference library.
Because ion-trap detectors measure all ions retained in the trapping steps, they do not experience sensitivity losses when run in the full-scan mode compared, for example, with triple-quadrupole MS (11). In this report, we describe an LC method coupled to ion-trap MS operated in full MS-MS scan mode to screen, identify, and quantify 8 sulfonylureas of interest in clinical toxicology.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
biological samples
Blank human plasma samples were supplied from the local blood bank (Etablissement Français du Sang, Reims, France). Authentic patient plasma samples had been submitted to our laboratory for toxicologic analysis.
calibration solution and calibration curve
A stock solutions of each sulfonylurea was prepared in methanol at a concentration of 1 g/L and stored at 4 °C. The stock solutions were further diluted with a mixture of 1 g/L formic acid (in purified water) and acetonitrile (50:50 by volume) to give a series of working solutions used to prepare the calibrators. Calibration curves were prepared by adding 10 µL of the appropriate calibrator to 0.5 mL of human blank plasma; the final concentrations in plasma were 3.9, 7.8, 15.6, 31.25, 62.5, 125, and 250 µg/L for glibenclamide; 7.8, 15.6, 31.25, 62.5, 125, 250, and 500 µg/L for glipizide, gliclazide, and glibornuride; 15.6, 31.25, 62.5, 125, 250, 500, and 1000 µg/L for glimepiride; 31.25, 62.5, 125, 250, 500, 1000, and 2000 µg/L for chlorpropamide and carbutamide; and 78.1, 156.25, 312.5, 625, 1250, 2500, and 5000 µg/L for tolbutamide.
sample preparation
Plasma samples (0.5 mL) were extracted with 2.5 mL of diethyl ether after addition of 25 µL of an internal standard (IS) solution [10 mg/L glisoxepide in 1 g/L formic acid–acetonitrile (50:50 by volume)] and 100 µL of 1 mol/L HCl. The mixture was vortex-mixed for 1 min, and then centrifuged at 3000g for 5 min. The organic layer was transferred to conical glass tubes and evaporated to dryness under a nitrogen stream at 40 °C. Finally, the residue was dissolved in 150 µL of 1 g/L formic acid–acetonitrile (50:50 by volume), and 5 µL was injected on the LC column.
lc-ms-ms
Instrumentation and chromatographic conditions.
The LC-tandem MS (MS-MS) system consisted of a ThermoFinnigan Surveyor® LC system equipped with an autosampler. Compounds were detected, identified, and quantified in plasma by use of a ThermoFinnigan LCQ Advantage® trap ion mass spectrometer linked to a ThermoFinnigan Xcalibur® data system. Chromatographic separations were carried out on a Hypurity C18 column [150 x 2.1 mm (i.d.); particle size, 5 µm; ThermoHypersil-Keystone] with the temperature maintained at 30 °C. Samples were eluted with a mobile phase consisting of acetonitrile and 1 g/L formic acid in purified water (50:50 by volume), delivered at a flow rate of 0.3 mL/min. During use, the mobile phase was degassed by the integrated Surveyor series degasser. The entire flow was directed into the ionization source without splitting. To optimize the MS-MS conditions and for creating library spectra, direct injection experiments were performed with a 500-µL syringe connected to a pump set at a flow rate of 5 µL/min. Appropriate solutions were prepared by dissolving the corresponding sulfonylureas in mobile phase to obtain a final concentration of 10 mg/L for each compound.
MS conditions.
The ionization technique used was electrospray ionization (ESI) in the positive-ion mode for glibenclamide, glipizide, gliclazide, glibornuride, glimepiride, and carbutamide and in the negative-ion mode for chlorpropamide and tolbutamide. The spray needle was set at a potential of 4 kV. The heated capillary was set at 200 °C, and the stainless-steel capillary was held at a potential of 10 V. Nitrogen was used as drying and nebulizing gas: Drying gas temperature was set at 300 °C, drying gas flow at 10 (arbitrary units), and nebulizing gas pressure at 240 kPa. The sheath gas flow rate of nitrogen was set at 40 (arbitrary units). The tube lens offset was set at 40 V and the electron multiplier voltage at 400 V peak to peak. Ultrapure helium (99.995%) was used in the trap as damping and collision gas (pressure of helium, 5.10–3 Torr). The instrument was set to acquire 3 microscans, and ion injection time into the trap was optimized by use of the integrated automatic gain control software.
MS conditions for detection, identification, and quantification.
Sulfonylureas were detected by LC-MS-MS in full MS-MS scan mode (m/z 150–600). Full-scan MS-MS spectra were obtained by collision-induced dissociation of each molecular ion with a normalized collision energy of 50%. To generate fragment ions of the molecular ion through collision-induced dissociation, 2 analytical runs (7 and 3 alternating scan events) were carried out at m/z 272, 494, 367, 324, 491, 446, and 450, which correspond to the protonated molecular ions [M+H]+ of carbutamide, glibenclamide, glibornuride, gliclazide, glimepiride, glipizide, and glisoxepide (IS), respectively, and at m/z 275, 269, and 448, which correspond to the deprotonated molecular ions [M–H]– of chlorpropamide, tolbutamide, and glisoxepide (IS), respectively. All full MS-MS spectra were recorded by scanning from m/z 150 to 600. Total run times were 7.5 min for positive- and 4 min for negative-ion mode analysis.
Reference MS-MS spectra of all sulfonylureas were collected individually by direct injection via the integrated syringe pump. These spectra, obtained by use of a normalized collision energy of 50%, were added to a custom full MS-MS library (including 
2000 compounds to date). Positive peaks were identified by comparing the underlying ESI mass spectra with the reference spectra in our MS-MS library.
Quantification was also performed in the full-scan MS-MS mode. Once full mass spectra of the product ions were generated, postacquisition data processing was designed to select particular ions for quantification (usually fragment ions with greater intensities). Peak-area ratios of the target ions of each sulfonylurea vs that of the IS (glisoxepide) were compared with calibration curves prepared under the same conditions. The latter were analyzed by unweighted linear regression. It should be noted that for chlorpropamide, tolbutamide, and carbutamide, the ranges of the calibration curves were significantly lower than their known therapeutic ranges (Table 3
). This was done mainly to allow the detection and quantification of very low drug concentrations as seen days or weeks after patients stopped their surreptitious use. If drug concentrations in authentic samples exceeded the calibration range, samples were reanalyzed after appropriate dilution with drug-free plasma.
|
method validation
Quality control.
Quality controls were prepared from a pool of blank human plasma with 3 different concentrations of each sulfonylurea added to give low, medium, and high concentrations (Table 1
). Plasma aliquots were stored at –20 °C until assayed and were renewed every 3 months.
|
Imprecision and recovery.
Imprecision and recovery were assessed by replicate analysis of 10 (intraday) and 20 (interday) quality-control samples over the 3 concentrations of each of the 8 sulfonylureas. Imprecision (as CV) was expected to be <15% except at the limit of quantification (LOQ; defined as the lowest concentration giving a signal-to-noise ratio >10:1), where 20% was acceptable. Recovery was calculated as: (mean measured concentration/added concentration) x 100. A recovery of 100 (15)% was considered acceptable, except at the LOQ, where 100 (20)% was acceptable.
Limits.
To determine the limit of detection (LOD; defined as the lowest concentration giving a signal-to-noise ratio >3:1), quality controls with decreasing amounts of each compound were assayed. Criteria for the LOQ were fulfilled by the lowest point of the calibration curve.
Carryover.
The lack of carryover effect was assessed by alternately analyzing blank plasma samples (n = 3) and plasma samples containing concentrations at the upper LOQ of each compound (n = 3). The residual concentration found in the first blank plasma sample following a high concentration sample was used to calculate the rate of carryover. It was considered minimal if <0.5% of the LOQ.
Extraction recoveries.
Extraction recoveries from human plasma were evaluated at low and high concentrations (n = 5). The samples were extracted without IS according to the procedure described above; 25 µL of the IS solution was then added to the organic phase, and the sample was evaporated to dryness. The residue was dissolved in 150 µL of mobile phase before analysis. For controls (n = 5), 25 µL of IS was added to mixtures of the 8 sulfonylureas prepared in mobile phase at the low and high concentrations; the mixtures were then gently evaporated. The residue of each calibration mixture was then dissolved in 150 µL of mobile phase and analyzed. Recoveries were calculated by comparing the peak areas of controls with those of plasma samples with added IS.
Specificity and ion suppression test.
We evaluated the specificity of the method by analyzing 10 different plasma samples obtained from healthy volunteers who had not received any of the sulfonylureas under investigation. The ion suppression effect for the method was also assessed with these plasma samples. After extraction, each plasma sample was injected into the LC-MS-MS system while high concentrations of the 8 drugs and IS were continuously infused post column (flow rate of 5 µL/min), as described by Müller et al. (12). Because glisoxepide and chlorpropamide elute at the same time in the negative-ion mode, additional ion suppression experiments were carried out to check for a potential mutual ion suppression effect of these 2 compounds. Briefly, extracted samples containing a low concentration of either glisoxepide or chlorpropamide, with or without a high concentration of the corresponding compound added, were injected into the LC-MS-MS system (n = 3).
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
. All full MS-MS spectra showed characteristic patterns, allowing unambiguous and rapid identification of the compounds by comparison with our full MS-MS reference library. For quantification purposes, glisoxepide was chosen as an IS because this compound has not been marketed in France. However, because the presence of glisoxepide in a patient’s sample can never be fully excluded, patient samples were extracted without IS added and analyzed before the final analysis. If one or several sulfonylureas were positively identified and confirmed, quantification of the corresponding compound(s) was performed in the full-scan MS-MS mode. Postacquisition data processing of full-scan MS-MS data permitted the "extraction" of analytes of interest by selection of specific product ions. This mode is known to be the most sensitive MS setup for an ion-trap detector; it thus permits quantitative analysis of analytes in complex matrices with high sensitivity. Indeed, because ion traps measure all ions retained in the trapping steps, they do not experience sensitivity losses in full-scan mode as does, for example, triple-quadrupole MS (11). Moreover, full-scan MS-MS mode provides additional specificity over conventional selected- or multiple-reaction monitoring experiments without sacrificing sensitivity.
chromatography
Reconstructed ion chromatograms of a blank plasma enriched with IS and therapeutic concentrations of all 8 sulfonylureas are shown in Fig. 1
. In the positive-ion mode, chromatographic separation is almost complete within 8 min; therefore, in light of the absence of coeluting compounds, no interference by ion suppression, particularly in the high concentration range, is expected. In the negative-ion mode, glisoxepide (IS) and chlorpropamide almost coeluted, and total run time was shorter than 4 min per sample.
|
validation data
The results of the method validation in human plasma are listed in Table 2
. Imprecision and recovery were within ranges defined as acceptable for bioanalytical purposes (13). Intraday imprecision (as CV) ranged from 1.8% to 18%, and recovery ranged from 81.3% to 118.2%. Interday imprecision did not exceed 10% over the 3 concentrations, except for the low concentration of tolbutamide (17%). The interday recovery of the method was satisfactory. Calibration curves were linear over the working concentration ranges with coefficients of determination (r2) >0.990 in all cases (Table 3
). The LOD are reported in Table 3
. LOQ were defined as the lowest concentration used for the calibration curves with a signal-to-noise ratio 
10 and for which the CV did not exceed 20%, and recoveries were 80%–100% (Table 3
). Under our experimental conditions, the carryover effect was minimal, with carryover <0.3% of the LOQ.
|
Extraction recoveries from human plasma were acceptable for glibenclamide, glibornuride, gliclazide, glimepiride, and glipizide with values ranging between 68% and 87% at low concentrations and between 63% and 83% at high concentrations (Table 3
). In contrast, carbutamide, chlorpropamide, and tolbutamide had low recoveries, most notably tolbutamide (<25%). Nevertheless, these recoveries were very reproducible at each concentration, and good calibration curves and coefficients of determination could be obtained. Moreover, the high sensitivity of the MS instrument made it possible to detect and quantify these 3 compounds from subtherapeutic to toxic concentrations (after appropriate dilution of samples if required) with good precision.
The analysis of 10 blank plasma samples from healthy volunteers showed no interfering peaks on the chromatograms. It is well known that ion suppression (which can be caused by interactions between matrix and analyte in solution when sprayed by the atmospheric pressure ionization source) is usually encountered with ESI (14); thus, this effect is usually evaluated when ESI spectrometry is used (12). In our assay, we observed no ion suppression effect in any of the blank plasma extracts at the expected retention times of the different sulfonylureas. In agreement with Müller et al. (12), we observed ion suppression effects at the LC solvent front (retention time <2 min, i.e., during the elution of nonretained compounds), but this did not interfere with the ionization of sulfonylureas assayed in either positive- or negative-ion mode. Additionally, this shows that appropriate chromatographic separation is needed before introduction of the sample into the ion source because we found no ion suppression for the analytes of interest in their specific retention-time windows. Other ion suppression experiments showed that no reciprocal ion suppression of glisoxepide and chlorpropamide occurred during their coelution.
application to authentic clinical cases
Between January 2003 and December 2004, we searched for the presence of sulfonylurea-type hypoglycemic drugs in 134 French patients who presented with unexplained and severe hypoglycemia. Using the presented analytical method, we found that 9 of these cases were likely to be related to surreptitious use of sulfonylureas [4 men with a mean (SD) age of 64.0 (11.8) years and 5 women with a mean age of 63.2 (23.3) years]. Glibenclamide, glimepiride, and gliclazide were detected alone in 3, 2, and 4 patients, respectively. As shown in Table 4
, plasma concentrations of glibenclamide and glimepiride were usually in the therapeutic ranges. Concerning gliclazide, in 1 case, the plasma concentrations were 2 times higher than the upper limit of the therapeutic range (i.e., 4000 µg/L). Another case had a plasma concentration in the therapeutic range. Finally, plasma concentrations in the last 2 cases were well below the lower limit of the therapeutic range (i.e., 250 µg/L). Because the elimination half-life of gliclazide in humans is 20 h, such low concentrations may be attributable to a delay (several days) between ingestion and blood sampling.
|
In conclusion, this method allows for the rapid screening and reliable identification of sulfonylureas in human plasma. Detection of sulfonylureas in plasma is performed by use of an ion-trap mass spectrometer with a highly selective LC-MS-MS procedure combined with library searching of MS-MS spectra. Quantification is achieved by use of a precise, accurate, and sensitive method developed in full-scan MS-MS mode. This assay has been successfully applied to the detection of Munchausen syndrome in authentic cases of patients with hypoglycemic crises of unknown origin.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
