Quantitative, Highly Sensitive Liquid Chromatography ... - CiteSeerX

3 downloads 71 Views 335KB Size Report
megestrol acetate, methylprednisolone, prednisolone, prednisone, triamcinolone, and triamcinolone acetonide. Methods: Stable isotopes of cortisol-9,11,12 ...
Clinical Chemistry 50:12 2345–2352 (2004)

Endocrinology and Metabolism

Quantitative, Highly Sensitive Liquid Chromatography–Tandem Mass Spectrometry Method for Detection of Synthetic Corticosteroids Robert L. Taylor, Stefan K. Grebe, and Ravinder J. Singh* Background: Measurements of serum or urine concentrations of synthetic glucocorticoids are useful for assessing suspected iatrogenic hypothalamic-pituitary-adrenal axis suppression and Cushing syndrome. We have developed a liquid chromatography–tandem mass spectrometry (LC-MS/MS) assay for the simultaneous quantitative analysis of beclomethasone dipropionate, betamethasone, budesonide, dexamethasone, fludrocortisone, flunisolide, fluorometholone, fluticasone propionate, megestrol acetate, methylprednisolone, prednisolone, prednisone, triamcinolone, and triamcinolone acetonide. Methods: Stable isotopes of cortisol-9,11,12,12-d4 and triamcinolone-d1 acetonide-d6 were added as internal standards to calibrators, controls, and unknown samples. After acetonitrile precipitation, these samples were extracted with methylene chloride, and the extracts were washed and dried. Reconstituted extract (15 ␮L) was injected on a reversed-phase column and analyzed by LC-MS/MS in positive-ion mode. Assay precision, accuracy, linearity, and sample stability were determined by use of enriched samples. Clinical validation included analysis of 8 serum and 20 urine samples from patients with undetectable cortisol concentrations and analysis of different types of tablets. Results: Functional assay sensitivity was as low as 0.6 –1.6 nmol/L for all compounds except for triamcinolone (7.6 nmol/L). Interassay CVs were 3.0 –20% for concentrations of 0.6 –364 nmol/L for all analytes. Recoveries of all analytes (except triamcinolone in serum) were 82–138% at 19.2– 693 nmol/L. All but one of the serum and urine samples from patients who were tested

Department of Laboratory Medicine & Pathology, Mayo Clinic and Foundation, Rochester, MN. *Address correspondence to this author at: Hilton 730, Department of Laboratory Medicine & Pathology, Mayo Clinic and Foundation, 200 First Street SW, Rochester, MN 55905. Fax 507-284-9758; e-mail singh.ravinder@ mayo.edu. Received April 27, 2004; accepted September 20, 2004. Previously published online at DOI: 10.1373/clinchem.2004.033605

because of suppressed cortisol concentrations contained at least one synthetic steroid. Tablet analysis recovered 75% of the synthetic steroids in suspected drugs. Conclusions: LC-MS/MS allows simultaneous quantitative detection of various synthetic steroids in serum, plasma, urine, and tablets. This provides a valuable tool for evaluating the clinical effects of topical and systemic synthetic corticosteroids. © 2004 American Association for Clinical Chemistry

Glucocorticoids are used for the treatment of a wide range of diseases. Unfortunately, their systemic use is often associated with significant side effects, ranging from skin fragility to full-blown iatrogenic Cushing syndrome (1– 4 ). Prolonged use may lead to suppression of the hypothalamic-pituitary-adrenal (HPA)1 axis. Consequently, there has been a push to replace systemic corticosteroids with topical preparations whenever possible. However, it is clear that this does not avoid all side effects. For example, injection of steroids into inflamed joints may be followed by systemic clinical improvement in distant, noninjected injected joints (5 ), and administration of inhaled corticosteroids is associated with dose-dependent HPA-axis suppression (6 –12 ). The systemic effects of topical corticosteroids can be so pronounced that clinicians occasionally see patients who display symptoms of Cushing syndrome but have undetectable cortisol concentrations and only a history of topical synthetic corticosteroid administration. Finally, HPA-axis suppression with or without symptoms may also be observed in some patients receiving nutritional supplements, herbal remedies, or alternative-medicine therapies that, unbeknownst to patients and physicians, contain synthetic corticosteroids.

1 Nonstandard abbreviations: HPA, hypothalamic-pituitary-adrenal; LCMS/MS, liquid chromatography–tandem mass spectrometry; PBS, phosphatebuffered saline; BSA, bovine serum albumin; and MRM, multiple reaction monitoring.

2345

2346

Taylor et al.: LC-MS/MS for Detection of Synthetic Corticosteroids

In all of these cases, establishing a definitive diagnosis of systemic synthetic corticosteroid effects and providing accurate monitoring require measurement of actual synthetic steroid concentrations in serum or urine. Unfortunately, such assays are generally not available in hospital, reference, and research laboratories, and if offered they may allow measurement of only a single synthetic corticosteroid. For endogenous steroids, methods based on liquid chromatography/electrospray tandem mass spectrometry have been developed. For synthetic steroids, assays have been developed specifically to analyze a single or only a few synthetic glucocorticoids and their metabolites (13–17 ). We therefore developed a liquid chromatography–tandem mass spectrometry (LC-MS/ MS) assay for the simultaneous detection and quantification of the 14 most frequently used synthetic corticosteroids in human serum, plasma, and urine and in tablet extracts.

Materials and Methods materials Beclomethasone dipropionate, betamethasone, budesonide, dexamethasone, fludrocortisone, flunisolide, fluorometholone, megestrol acetate, methylprednisolone, prednisolone, prednisone, triamcinolone, and triamcinolone acetonide were purchased from Sigma. Methanol and methylene chloride were HPLC grade (EM Science). A 1.92–2.79 mmol/L (1 g/L) stock solution of these analytes was prepared in methanol. A 19.2–27.9 ␮mol/L (10 mg/L) working solution was prepared by diluting the stock solution 1:100 with methanol–water (70:30 by volume) containing 4 ␮mol/L (1 mg/L) estriol. The estriol in the reconstitution solvent was added to prevent loss of the extracted analytes by nonspecific binding to the glass surface. To prepare the fluticasone propionate working standard, we obtained a 0.1 ␮mol (50 ␮g)/spray metered dose inhaler of Flonase® from the pharmacy. Three sprays were directed to waste followed by two sprays (0.2 ␮mol; 100 ␮g) into a glass tube. The contents of the glass tube were quantitatively transferred to a 10-mL volumetric flask by use of a 700 mL/L methanol solution, creating a 20 ␮mol/L (10 mg/L) working solution. The process was repeated three times. Cortisol-9,11,12,12-d4 was purchased from Cambridge Isotope Laboratories (isotopic enrichment, 98%). Triamcinolone-d1 acetonide-d6 was purchased from CDN Isotopes (stated isotopic enrichment, 74% d1, 99% d6). We prepared a stock solution containing 2.73 mmol/L cortisol-9,11,12,12-d4 and 2.27 mmol/L triamcinolone-d1 acetonide-d6 (1 g/L of each internal standard) in methanol. A working solution containing 11 ␮mol/L cortisol9,11,12,12-d4 and 9.1 ␮mol/L triamcinolone-d1 acetonide-d6 (4 mg/L) was prepared by diluting the stock solution 1:250 with methanol–water (70:30 by volume) containing estriol.

sample preparation Tablet extracts required additional processing, which was not necessary for serum, plasma, or urine samples. We prepared tablet extracts by weighing the tablet, recording the weight, and then crushing the tablet into powder with a mortar and pestle. We weighed a portion of the powder and added it to a 10-mL volumetric flask. Methanol was added to the 10-mL volume, and the sample was mixed with a magnetic stirrer for 1 h at room temperature. A four-point calibration curve [0 ␮g/L, 5 ␮g/L (9.6 –14 nmol/L), 50 ␮g/L (96 –140 nmol/L), and 250 ␮g/L (480 – 700 nmol/L)] in phosphate-buffered saline containing 10 g/L bovine serum albumin (PBS-BSA) and two concentrations of controls were run with each assay. Aliquots of urine, tablet extract, serum, or plasma were centrifuged for 5 min at 1000g to remove particulate matter. Tablet extracts were diluted 1:10, 1;100, and 1:1000 in PBS-BSA. We transferred 0.5-mL fractions of the calibrators, controls, urine samples, serum samples, plasma samples, and diluted tablet extracts to 12 ⫻ 75 mm borosilicate glass tubes and mixed them with 25 ␮L of the working solution containing 11 ␮mol/L cortisol-9,11,12,12-d4 and 9.1 ␮mol/L triamcinolone-d1 acetonide-d6 and 0.5 mL of acetonitrile. We then centrifuged these solutions at 1000g for 10 min. The supernatants were transferred to 13 ⫻ 100 mm borosilicate glass tubes, and 4 mL of methylene chloride was added to each tube. The samples were then vortex-mixed for 30 s and centrifuged at 1000g for 5 min; the upper aqueous layer was then aspirated and discarded. The methylene chloride fractions were washed sequentially with 1.0 mL of 0.1 mol/L sodium hydroxide, 1.0 mL of 0.1 mol/L hydrochloric acid, and 1.0 mL of H2O, and each aqueous layer was aspirated and discarded. The washed methylene chloride was evaporated under nitrogen at 45 °C, and the dried extract reconstituted with 75 ␮L of methanol–water (70:30 by volume) containing estriol. The reconstituted extracts were centrifuged for 5 min at 1000g and transferred to autosampler vials.

lc-ms/ms All LC-MS/MS experiments were performed with PerkinElmer Series 200 pumps and autosampler for sample introduction and an API 3000TM tandem mass spectrometer (Applied Biosystems), operating with an electrospray ionization source. The analytes and internal standards were chromatographically resolved from other sample components with a reversed-phase column [SYNERGI 4␮ MAX-RP; 50 ⫻ 4.6 mm (i.d.); Phenomenex®] combined with a 4 ⫻ 2 mm (i.d.) precolumn filter of the same material. The mobile phase, delivered at a flow rate of 750 ␮L/min, consisted of 29% acetonitrile in 0.1 mmol/L ammonium acetate for 9.5 min, followed by 65% acetonitrile in 0.1 mmol/L ammonium acetate for 4 min, with a final 0.5-min reequilibration at 29% acetonitrile in 0.1 mmol/L ammonium acetate. The flow delivered to the TurboIonspray® was split, with 80% going to waste and 20% to the ionization probe. Total instrument analysis

Clinical Chemistry 50, No. 12, 2004

time, including sample introduction and sample run time, was 15 min per sample. Urine, serum, and plasma results were calculated directly from the calibration curve. After the values for the tablet extracts were calculated from the calibration curve, the amount of analyte per tablet was calculated as follows: Value (␮g/L) ⫻ dilution factor ⫻ 0.1 L (flask volume) ⫻

total tablet weight ⫽ ␮g/tablet powder weight of tablet used

The mass spectrometer operating conditions consisted of a source heater probe of 450 °C, with a TurboIonspray voltage of 5000 V, declustering potential of 35, entrance potential of 10, cell exit potential of 12, nebulizer gas setting of 8, curtain gas setting of 6, and CAD setting of 8. Collision energies were different for several analytes and are listed in Table 1. Data acquisition and quantitative processing were accomplished using AnalystTM software, (Applied Biosystems). The ions for all glucocorticoids were generated in the positive-ion mode. The ion transitions used to monitor the analytes were determined by infusing 10 mg/L solutions of each analyte separately in methanol–water (50:50 by volume) containing 0.2 mmol/L ammonium acetate at very low flow rate of 10 ␮L/min. The multiple-reaction monitoring (MRM) ion transitions for the synthetic steroids are listed in Table 1.

method validation Method validation included extraction efficiency, precision, functional assay sensitivity, recovery, linearity, and stability. Pools containing different analyte concentrations Table 1. MRM ion transitions and retention times for synthetic glucocorticoids. m/z, amua Analyte

Q1

Q3

Retention time, min

Collision energy, V

Beclomethasone dipropionate Betamethasone Budesonide Cortisol Cortisol-d4 Cortisone Dexamethasone Flunisolide Fluorometholone Fludrocortisone Fluticasone propionate Megestrol acetate Methylprednisolone Prednisolone Prednisone Triamcinolone Triamcinolone acetonide Triamcinolone-d1 acetonide-d6

521 393 431 363 367 361 393 435 377 381 501 385 375 361 359 395 435 442

337 147 413 121 121 163 147 321 279 239 239 224 161 147 147 225 397 404

13.2 5.56 11.1 3.15 3.15 3.57 5.92 8.81 10.7 3.46 12.6 12.8 4.90 3.02 3.30 1.94 8.22 8.07

22 37 22 37 37 37 37 22 37 37 22 37 37 37 37 37 22 22

a

amu, atomic mass units.

2347

for the method validation studies were created by adding various amounts of the calibrators to normal urine and serum and dividing them into aliquots, which were stored at ⫺20 °C. Extraction efficiencies were determined by comparing a 15-␮L injection of an unextracted 200 ␮g/L (392–560 nmol/L) calibrator vs a 15-␮L injection of the 50 ␮g/L (98 –140 nmol/L) PBS-BSA-extracted calibrator, which was also concentrated four times. Intraassay precision studies included 20 samples for each concentration. Interassay imprecision was calculated based on 23 separate assays, run on separate days. All samples were assayed in singlet. The functional assay sensitivity for each analyte was defined as the lowest analyte concentration with an interassay CV ⬍20%. For the recovery studies, we added all 14 analytes at three concentrations each, ranging from 19.2 to 693 nmol/L (10 –250 mg/L), to urine, serum, EDTA-plasma, and heparin-plasma samples from three patients that were devoid of synthetic corticosteroids. Recoveries are expressed as the percentages of added analytes that were recovered. To determine assay linearity, we added 250 ␮g/L (480 – 693 nmol/L) each of the calibrators to three patient samples for each of the sample matrices and diluted them 2-, 5-, and 20-fold in the PBS-BSA buffer. The expected value of each dilution was calculated based on the result of the undiluted sample. We evaluated assay linearity by dividing the observed value by the expected value for each dilution, with the result being expressed as the percentage of expected recovery. For stability studies, we added calibrators to five patient samples each of urine (unpreserved, boric acid, and acetic acid), serum (SST and clot tube), EDTA plasma, and heparin plasma. Four aliquots were immediately frozen at ⫺20 °C and subjected to zero, one, two, or three freeze–thaw cycles. In addition, one aliquot was stored at room temperature, and one aliquot was stored refrigerated. On days 1, 3, and 7, an aliquot was taken from the room temperature and the refrigerated samples and was frozen at ⫺20 °C. All aliquots were then assayed. To study the matrix effect of the assay, we prepared a 1 mg/L (1920 –2772 nmol/L) solution containing all of the synthetic steroids. We then used a T-valve inserted in the solvent line after the autosampler to infuse each solution into the system at 10 ␮L/min. While the solution was being infused, an extracted sample of zero calibrator in BSA (data not shown), a negative serum sample extract, or a negative urine extract was injected. We performed a limited Institutional Review Boardapproved clinical assay validation by retrospectively studying 8 plasma samples and 20 urine samples with undetectable cortisol concentrations. We also tested the dexamethasone content of two different types of dexamethasone tablets manufactured outside the United States with a manufacturer-stated drug content of 750

2348

Taylor et al.: LC-MS/MS for Detection of Synthetic Corticosteroids

mg/tablet. Finally, we analyzed another tablet, of unknown origin, that had been retrieved from a patient who had undetectable serum cortisol concentrations but exhibited symptoms of Cushing syndrome.

Results lc-ms/ms method characteristics With the LC-MS/MS method, the signal intensities for all of the steroids were better in positive mode compared with the negative mode because of better electrospray ionization of positively charged steroids. The LC-MS/MS method was optimized to detect these steroids, using the MRM pairs comprising the precursor and product ions. The daughter ions (Q3) of all of the steroids obtained during collision of the precursor ions (Q1) in tandem MS are listed in the Table 1. The structures of the various synthetic steroids (see the Data Supplement that accompanies the online version of this article at http://www. clinchem.org/content/vol50/issue12/) indicate that most of these steroids, except for triamcinolone, fludrocortisone, and prednisolone, are more hydrophobic than cortisol and thus elute later on a reversed-phase column with an increasing gradient of organic solvent (Table 1). The extracted sample ion chromatogram for a calibrator containing 250 ␮g/L (480 –963 nmol/L) each of the 14 different synthetic glucocorticoids and 2 internal standards is shown in Fig. 1. We would have preferred to use individual internal standards for all of the steroids studied, but only deuterium-labeled cortisol and triamcinolone acetonide were

Fig. 1. MRM chromatogram of a mixture of various synthetic steroids at a concentration of 250 ␮g/L in a calibrator. The order of the elution of these compounds is as follows; peak 1, triamcinolone; peak 2, prednisolone; peaks 3 and 4, cortisol and cortisol-d4; peak 5, prednisone; peak 6, fludrocortisone; peak 7, cortisone; peak 8, methylprednisolone; peak 9, betamethasone; peak 10, dexamethasone; peak 11, triamcinolone-d1 acetonide-d6; peak 12, triamcinolone acetonide; peak 13, flunisolide; peak 14, fluorometholone; peak 15, budesonide; peak 16, fluticasone propionate; peak 17, megestrol acetate; peak 18, beclomethasone dipropionate.

commercially available. On the basis of the chromatography and retention times of the steroids, we used cortisol-d4 for quantification of betamethasone, dexamethasone, fludrocortisone, methyl prednisolone, prednisolone, prednisone, triamcinolone, cortisol, and cortisone. For quantification of the rest of the steroids, we used the deuterium-labeled triamcinolone acetonide as internal standard. The peaks for the isomers dexamethasone and betamethasone have identical ion pairs and thus need to be separated chromatographically (Table 1 and Fig. 1). Similarly, the isomers triamcinolone acetonide and flunisolide also need to be separated chromatographically from each other because of similar precursor and product ions. Triamcinolone acetonide and flunisolide had similar parent (Q1) and product ions (Q3) of m/z 397 and 321. The m/z 397 ion was optimum for triamcinolone acetonide quantification, and we chose the m/z 321 ion for flunisolide because this daughter ion had a higher signal intensity. These compounds were separated by use of a slow gradient, and attention was also paid to relative retention times to ensure proper identification.

method validation The intraassay imprecision (CV) for all analytes ranged from 2.6% to 9.8% at 19.2–327 nmol/L (10 –118 ␮g/L). The interassay imprecision for all analytes ranged from 3.2% to 20% for mean concentrations of 0.6 –364 nmol/L (0.3– 130 ␮g/L) and are summarized in Table 2. From the imprecision data for the serum and urine matrices, we calculated that the functional sensitivity was 0.6 –1.6

Clinical Chemistry 50, No. 12, 2004

Table 2. Interassay precision. Level Ia

CV, %

Analyte

CV, %

S/Nb

Level II

Level III

Beclomethasone dipropionate Betamethasone Budesonide Dexamethasone Fludrocortisone Flunisolide Fluorometholone Fluticasone propionate Megestrol acetate Methylprednisolone Prednisolone Prednisone Triamcinolone Triamcinolone acetonide

19 19 17 19 20 14 19 20 17 18 12 20

10 4.3 7.9 4.0 5.5 7.7 15 9.5 39 19 32 7.6

13

33

16 10 13 9.3 17 12 14 16 12 9.3 12 11 15 6.2

9.4 9.6 9.2 9.7 9.1 3.5 16 11 9.2 11 8.6 8.0 9.9 3.2

Level I, 0.6 –1.6 nmol/L (0.3– 0.7 ␮g/L); level II, 4.4 –11.2 nmol/L (2.3– 4.0 ␮g/L); level III, 198 –363 nmol/L (103–130 ␮g/L). b S/N, signal-to-noise ratio. a

nmol/L (0.3– 0.7 ␮g/L) for all analytes except for triamcinolone. Triamcinolone had a functional sensitivity of 7.6 nmol/L (3 ␮g/L) because of poor recovery in the extraction protocol used in the present study, in which the goal was to extract numerous synthetic steroids of variable hydrophilicity and hydrophobicity. The signal-to-noise ratios for the analytes in the functional sensitivity pool were calculated and are listed in Table 2. The absolute extraction efficiency ranged from 69% to 89% for all analytes except triamcinolone, which had an extraction efficiency of 33%, presumably because of its more hydrophilic nature. The recovery data relative to the internal standards for various glucocorticoids are summa-

2349

rized in Table 3. Recoveries ranged from 82% to 138% (mean, 108%) for all analytes except triamcinolone. Triamcinolone had acceptable recovery in urine (range, 90 – 134%; mean, 110%), but recoveries in serum and plasma were highly variable (range, 28 – 80%; mean, 46%). This is probably because triamcinolone is the most polar component of the steroid profile and is lost during the extraction process. Although the method lacked deuterium-labeled triamcinolone as an internal standard, it is also likely that components of the serum matrix may have suppressed the signal and accounted for the variable recovery of this compound (18 ). The matrix effects of a negative serum sample extract and a negative urine sample extract are shown in the top and bottom panels of Fig. 2, respectively; we observed no significant matrix effects, probably because in our method, plasma and urine samples are not only extracted with a nonpolar solvent (methylene chloride) but are also washed with acid and base to remove most of the suppressing and interfering matrix components. With the exception of triamcinolone in serum, the method was linear, based on a recovery of 80 –120%, for all twofold dilutions. Using the criterion of ⫾20% recovery, all analytes except beclomethasone dipropionate were stable in serum, plasma, and urine for 1 day at ambient temperature (stable in urine for 7 days), for 7 days when stored refrigerated, and for three freeze–thaw cycles. Beclomethasone dipropionate was stable in frozen serum, plasma, and urine for three freeze–thaw cycles and in refrigerated urine for 7 days. The chromatogram of a patient sample with suppressed plasma cortisol and confirmed to be positive for methylprednisolone and megestrol acetate is shown in Fig. 3. Among eight patient samples, five were found to be

Table 3. Recovery of analytes added to urine, serum, and plasma. Recovery, %

Analyte

Beclomethasone dipropionate Betamethasone Budesonide Dexamethasone Fludrocortisone Flunisolide Fluorometholone Fluticasone propionate Megestrol acetate Methylprednisolone Prednisolone Prednisone Triamcinolone Urine Serum Triamcinolone acetonide

1 ␮g/L (190–280 nmol/L) added

5 ␮g/L (980–1400 nmol/L) added

20–25 ␮g/L (3840–6980 nmol/L) added

82–118 99–113 86–123 85–117 102–136 103–136 104–138 83–118 85–123 100–118 102–119 94–127

94–109 95–112 92–117 91–115 106–128 106–123 99–127 97–117 101–121 98–113 99–115 100–131

85–115 98–109 87–117 92–106 97–120 97–106 96–112 94–116 94–124 99–107 97–106 96–124

97–134 28–64 116–132

96–127 36–80 105–124

90–124 38–63 98–112

2350

Taylor et al.: LC-MS/MS for Detection of Synthetic Corticosteroids

Fig. 3. Chromatogram of synthetic steroids confirmed in a patient serum sample. The serum sample was found to be positive for methylprednisolone and megestrol acetate.

exhibiting symptoms of Cushing syndrome with undetectable cortisol concentrations, was found to contain 2.9 mg (7360 nmol)/tablet of triamcinolone.

Discussion Fig. 2. Effect of specimen type matrices on the signal intensities of various synthetic steroids in the LC-MS/MS method. A solution containing all of the synthetic steroids [1 mg/L (1920 –2772 nmol/L)] was fed into the solvent line through a T-valve after the autosampler and infused into the system at a flow rate of 10 ␮L/min. While the solution was being infused, an extracted sample of zero calibrator in BSA (data not shown), a negative serum sample extract (top), or a negative urine extract (bottom) was injected. To simplify the chromatogram, the effects of the above matrices on the signals of only five analytes (prednisolone, betamethasone, dexamethasone, triamcinolone acetonide, and megestrol acetate) are displayed.

positive for dexamethasone [1.5–28 ␮g/L (3.8 –71.4 nmol/ L)], one for megestrol acetate [23 ␮g/L (60 nmol/L)], and one for prednisone [2 ␮g/L (5.6 nmol/L)] and prednisolone [16 ␮g/L (44.4 nmol/L)], and one sample was negative. Of the 20 urine samples, 7 were positive for dexamethasone [2–26 ␮g/L (5– 66 nmol/L)], 5 for triamcinolone acetonide [0.8 –20 ␮g/L (2– 46 nmol/L)], 3 for methylprednisolone [0.8 – 6.5 ␮g/L (2–17 nmol/L)], 1 for prednisone [12 ␮g/L (34 nmol/L)] and prednisolone [1.8 ␮g/L (5 nmol/L)], 1 for dexamethasone [17 ␮g/L (43 nmol/L)] and triamcinolone acetonide [22 ␮g/L (51 nmol/L)], and 1 for triamcinolone acetonide [3.4 ␮g/L (8 nmol/L)] and budesonide [2.4 ␮g/L (6 nmol/L)]; 1 sample was negative. The tablets obtained from patients who had purchased them from outside the United States were found to contain 565 ␮g (1441 nmol)/tablet of dexamethasone. Another tablet of unknown origin, from a patient

HPLC-MS/MS methods are now commonly used in clinical laboratories for the analysis of various compounds, including steroids and biogenic amines, requiring very high sensitivity (19 –23 ). The specificity and sensitivity of the analysis have improved the diagnostic efficacy for Cushing syndrome and pheochromocytoma. The diagnosis of Cushing syndrome is complicated by iatrogenic or unwitting exposure to various synthetic steroids. Our method was developed specifically for screening of synthetic glucocorticoids in patients with symptoms of Cushing syndrome with low cortisol concentrations. Previously, methods for the detection of a limited number of steroids by MS have been reported (13–16 ). The method described here provides a good screening tool for patients with unexplained cortisol suppression or suspected iatrogenic Cushing syndrome. In this method, 14 different synthetic corticosteroids can be detected and quantified simultaneously. The assay also allows identification of these synthetic glucocorticoids in medications of unknown composition. This method can be adapted to incorporate new synthetic glucocorticoids and performed on serum, plasma, or urine. The synthetic glucocorticoids are sufficiently stable to allow testing on previously collected samples that have been properly stored. The initial protocol for this method used only the methylene chloride extraction and washes for sample preparation. This provided accurate recoveries for all

2351

Clinical Chemistry 50, No. 12, 2004

compounds except for megestrol acetate and fluticasone propionate in serum and plasma, which were unacceptably low at 15–50%. Acetonitrile pretreatment improved the recoveries of megestrol acetate and fluticasone propionate in serum and plasma. It may be that these two compounds are protein-bound in the serum and plasma, which may have decreased the recovery without the acetonitrile pretreatment. Analysis of a limited number of patient samples suggests that the proposed method is sufficiently sensitive to provide clinically valuable data to distinguish endogenous hypocortisolism from synthetic glucocorticoid-induced HPA-axis suppression and to allow diagnosis of iatrogenic Cushing syndrome. Among the plasma samples with undetectable cortisol, 87.5% were positive for at least one and 95% of the urine samples were positive for one or more synthetic corticosteroids. Because some of the synthetic steroids are extremely potent, depot injections into joints may release these potent substances for prolonged periods of time. Therefore, HPA-axis suppression or iatrogenic Cushing syndrome may be observed many weeks or months after topical synthetic steroid application. Indeed, during the validation phase of the assay, we received requests from clinical colleagues to investigate several such situations. As we have reported recently, we were able to detect previous synthetic steroid use in all of these cases, sometimes long after the drugs had been discontinued (23 ). Our assay could also be valuable for identifying individual patients who are at particular risk of developing corticosteroid side effects as a consequence of atypically large fractional systemic absorption of topical synthetic corticosteroids. Finally, the ability to detect these compounds at low concentrations is likely to improve pharmacokinetic, drug validation, and toxicity studies. The assay’s performance characteristics are for the most part very satisfactory. Unlike the other synthetic corticosteroids, the results for triamcinolone in serum are only semiquantitative. If needed, improvements in the extraction methodology could provide better quantitative triamcinolone measurements. For tablet analysis, the analyzed result was 75% of the manufacturer-stated amount of dexamethasone in the tablets, indicating that this method can be used for detection of synthetic glucocorticoids in unknown medications. The assay run time, although acceptable, is relatively long for a LC-MS/MS method. However, this is necessary to chromatographically separate the isomers dexamethasone and betamethasone and the isomers triamcinolone acetonide and flunisolide and to elute the more hydrophobic compounds in the profile. The step gradient increase of acetonitrile from 29% to 65% at 9.5 min elutes the more hydrophobic compounds in a reasonable time. The major benefit of this step gradient from hydrophilic mobile phase to hydrophobic phase is that new synthetic glucocorticoids may be added to the assay without altering the chromatographic conditions, which already allow

separation of steroids with a broad range of hydrophobicity. In addition, when interest is limited to one specific type of synthetic corticosteroid, rather than a profile, run conditions can be adjusted to allow a much more rapid analysis.

References 1. Adams N, Bestall JM, Jones PW. Fluticasone versus beclomethasone or budesonide for chronic asthma. Cochrane Database Systematic Reviews, 2002: CD002310. 2. Ganguly S, Dutta S, Chakraborty S, Ghosh J. Iatrogenic Cushing’s syndrome. Indian Pediatr 1998;35:1236 –7. 3. Clevenbergh P, Corcostegui M, Gerard D, Hieronimus S, Mondain V, Chichmanian RM, et al. Iatrogenic Cushing’s syndrome in an HIV-infected patient treated with inhaled corticosteroids (fluticasone propionate) and low dose ritonavir enhanced PI containing regimen. J Infect 2002;44:194 –5. 4. Skov M, Main KM, Sillesen IB, Muller J, Koch C, Lanng S. Iatrogenic adrenal insufficiency as a side-effect of combined treatment of itraconazole and budesonide. Eur Respir J 2002;20: 127–33. 5. Neidel J. Intra-articular steroid therapy for inflammatory rheumatic diseases in children and adolescents. Orthopade 2002;31:1175– 8. 6. Nelson HS, Stricker W, Casale TB, Raff H, Fourre JA, Aron DC, et al. A comparison of methods for assessing hypothalamic-pituitaryadrenal (HPA) axis activity in asthma patients treated with inhaled corticosteroids. J Clin Pharmacol 2002;42:319 –26. 7. Casale TB, Nelson HS, Stricker WE, Raff H, Newman KB. Suppression of hypothalamic-pituitary-adrenal axis activity with inhaled flunisolide and fluticasone propionate in adult asthma patients. Ann Allergy Asthma Immunol 2001;87:379 – 85. 8. Boner AL. Effects of intranasal corticosteroids on the hypothalamic-pituitary-adrenal axis in children. J Allergy Clin Immunol 2001; 108:S32–9. 9. Gupta D, Behera D, Lalrinmawia H, Dash RJ. Hypothalamopituitary-adrenal axis function in asthmatics taking low dose inhaled beclomethasone dipropionate. J Assoc Phys India 2000; 48:682– 4. 10. Jarvis B, Faulds D. Inhaled fluticasone propionate: a review of its therapeutic efficacy at dosages ⱕ500 ␮g/day in adults and adolescents with mild to moderate asthma. Drugs 1999;57:769 – 803. 11. Dluhy RG. Clinical relevance of inhaled corticosteroids and HPA axis suppression. J Allergy Clin Immunol 1998;101:S447–50. 12. Alsaeedi A, Sin DD, McAlister FA. The effects of inhaled corticosteroids in chronic obstructive pulmonary disease: a systematic review of randomized placebo-controlled trials. Am J Med 2002; 113:59 – 65. 13. Arthur KE, Wolff JC, Carrier DJ. Analysis of betamethasone, dexamethasone and related compounds by liquid chromatography/electrospray mass spectrometry. Rapid Commun Mass Spectrom 2004;18:678 – 84. 14. Guan F, Uboh C, Soma L, Hess A, Luo Y, Tsang DS. Sensitive liquid chromatographic/tandem mass spectrometric method for the determination of beclomethasone dipropionate and its metabolites in equine plasma and urine. J Mass Spectrom 2003;38: 823–38. 15. Noben JP, Gielen B, Royackers E, Missotten M, Jacobs A, Raus J. A high-performance liquid chromatography/tandem mass spectrometric screening method for eight synthetic corticosteroids in bovine feces and the simultaneous differentiation between dexamethasone and betamethasone. Rapid Commun Mass Spectrom 2002;16:1590 – 4. 16. Van Den Hauwe O, Perez JC, Claereboudt J, Van Peteghem C.

2352

Taylor et al.: LC-MS/MS for Detection of Synthetic Corticosteroids

Simultaneous determination of betamethasone and dexamethasone residues in bovine liver by liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 2001;15: 857– 61. 17. Antignac JP, Le Bizec B, Monteau F, Poulain F, Andre F. Collisioninduced dissociation of corticosteroids in electrospray tandem mass spectrometry and development of a screening method by high performance liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 2000;14:33–9. 18. Annesley TM. Ion suppression in mass spectrometry. Clin Chem 2003;49:1041– 4. 19. Lagerstedt SA, O’Kane DJ, Singh RJ. Measurement of plasma free metanephrine and normetanephrine by liquid chromatography– tandem mass spectrometry for diagnosis of pheochromocytoma. Clin Chem 2004;50:603–11.

20. Nelson RE, Grebe SK, O’Kane DJ, Singh RJ. Liquid chromatography–tandem mass spectrometry assay for simultaneous measurement of estradiol and estrone in human plasma. Clin Chem 2004;50:373– 84. 21. Beer PM, Bakri SJ, Singh RJ, Liu W, Peters GB III, Miller M. Intraocular concentration and pharmacokinetics of triamcinolone acetonide after a single intravitreal injection. Ophthalmology 2003;110:681– 6. 22. Taylor RL, Machacek D, Singh RJ. Validation of a high-throughput liquid chromatography-tandem mass spectrometry method for urinary cortisol and cortisone. Clin Chem 2002;48:1511–9. 23. Kumar S, Singh RJ, Reed AM, Lteif AN. Cushing’s syndrome after intra-articular and intradermal administration of triamcinolone acetonide in three pediatric patients. Pediatrics 2004;113: 1820 – 4.