| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT 84108
2 Departments of Medicine and Pathology, University of Utah, Salt Lake City, UT 84132
aauthor for correspondence: fax 801-584-5207, e-mail kushnmm{at}aruplab.com
For the analysis of urinary free cortisol (UFC), liquid chromatography (LC), LC-mass spectrometry (LC-MS), and liquid chromatography-tandem MS (LC-MS/MS) offer better specificity and accuracy of result than immunoassay-based methods (1). The high specificity of LC-MS/MS provides simplification of sample preparation, reduced costs, increased sample throughput, and a low rate of sample interference. Published approaches to LC-MS/MS analysis of UFC use either sample extraction (1) or direct sample injection (2). The direct methods reduce the labor requirements and decrease the potential for human error. We have developed a direct sample injection method similar to one published by Nassar et al. (2). The primary changes in the proposed method are the use of a guard cartridge for trapping cortisol as part of an online purification strategy and the use of step gradient rather than isocratic LC conditions. Both electrospray ionization (ESI) and atmospheric pressure ionization (APCI) are well suited for LC-MS/MS. We used APCI because preliminary experiments demonstrated a two- to threefold increase in sensitivity compared with ESI. Better sensitivity for cortisol with an APCI interface compared with ESI can be explained by the poor ionization of cortisol in solution.
Cortisol and ammonium formate were purchased from Sigma, and d4-cortisol was purchased from Cambridge Isotope Laboratories. Methanol (HPLC grade) was purchased from Fisher Scientific. Patient samples were analyzed for cortisol within 1–3 days after collection. All studies with samples from humans were approved by the Institutional Review Board of the University of Utah.
Sample preparation for the method was performed as follows. A 500-µL aliquot of centrifuged urine sample and 100 µL of the working internal standard (IS; d4-cortisol, 0.5 mg/L) were added to autosampler vials, and the vials were vortex-mixed. A PE series 200 HPLC system (Perkin-Elmer Analytical Instruments) was equipped with a Luna C18 column [50 x 2.0 mm (i.d.); 5 µm particles; Phenomenex]. Sample trapping and purification was performed online with a C18 guard cartridge (Phenomenex). Urine with IS was injected with mobile phase consisting of methanol–5 mmol/L ammonium formate (7:93 by volume). The cartridge was washed for 0.5 min to remove hydrophilic constituents. After 0.5 min, the mobile phase composition was changed by step gradient to 55:45 methanol–5 mmol/L ammonium formate, and the column effluent was directed to the LC column connected to a MS interface. The column temperature was 40 °C, the injection volume was 200 µL, and the injection interval was 8 min. The autosampler syringe wash solvent was methanol–water (80:20 by volume) containing 8 mmol/L trifluoroacetic acid.
An API 2000 (Applied Biosystems/MDS SCIEX) tandem mass spectrometer was used in the positive-ion mode with an APCI interface. Quantitative analysis was performed in the multiple-reaction monitoring mode, and the transitions monitored were m/z 363
121 and 36397 for cortisol and m/z 367121 and 36797 for d4-cortisol. Quantification was performed based on the m/z 121 product ion, whereas the m/z 97 product ion was used for confirmation of cortisol and d4-cortisol identity. The collision gas was nitrogen with a collision cell pressure of 1.1 Pa. The APCI needle current was 2.0 µA, the orifice voltage was 60 V, and the collision energy was 24 V for the m/z 363121 and 367121 transitions and 52 V for the m/z 36397 and 36797 transitions. Quantitative data analysis was performed with TurboQuanTM (Applied Biosystems/MDS SCIEX) software.
The calibrators were prepared at 10, 50, 100, and 200 µg/L in a solution prepared by dissolving 16 mmol of potassium hydrogen phosphate, 5 mmol of sodium dihydrogen phosphate, 60 mmol of sodium chloride, 340 mmol of urea, and 9 mmol of creatinine in 1 L of deionized water. The 50 µg/L calibrator was used to establish a qualitative intensity ratio of the product ion fragments for cortisol and d4-cortisol. For cortisol, the ratio was the intensity of the m/z 36397 transition divided by the intensity of the m/z 363121 transition. For d4-cortisol, the ratio was the intensity of the m/z 36797 transition divided by the intensity of the m/z 367121 transition. The acceptability limits for qualitative ion intensity ratio for the controls and test samples were set at ±40% of the values established with the calibrator. Monitoring of at least two MS/MS transitions allows detection of interferences. A ratio above the allowable upper limit indicated interference at the m/z 363121 transition, in which case quantification was performed using the m/z 36397 transition.
Within-run imprecision of triplicate analyses of human urine samples supplemented with cortisol to concentrations of 10, 43, and 75 µg/L was <5.8%, between-run imprecision for the samples analyzed over 5 days was <6.5%, and total imprecision was <8.5%. The limit of quantification (2 µg/L) was determined as the lowest concentration at which accuracy was within ±20%, imprecision (CV) was <15%, and the qualitative ion ratio was within the limits established by a calibration performed with every run. The limit of detection, defined as the lowest concentration at which the cortisol peak was detected with signal-to-noise ratio >3 and retention time was consistent with cortisol, was 1 µg/L. The calibration curve was linear up to 8000 µg/L.
We compared results of UFC analyses in 18 patient samples containing UFC at concentrations of 2–560 µg/L by three methods: (a) the evaluated method; (b) a RIA (Diagnostic Products Corporation) (3); and (c) an extraction-based LC-MS/MS method (1). The slope of linear regression for agreement with the RIA method showed a positive proportional bias of 56% compared with the HPLC-MS/MS method with direct sample injection, similar to the 
53% positive proportional bias reported for a chemiluminescence immunoassay compared with an extraction-based HPLC-MS/MS method (1). Such a difference is likely attributable mainly to cross-reactivity in the immunoassay-based methods. The linear regression equation for the proposed method (y) and the comparison HPLC-MS/MS method (x) was: y = (0.99 ± 0.003)x - (1.24 ± 0.50) µg/L; r = 0.999; Sy|x = 7.12 µg/L; t-test value, 0.61 (tcrit at 99% confidence is 2.72).
We observed an unidentified interfering peak at the m/z 363121 transition in 1% of tested patient samples. The interfering peak eluted at a retention time slightly shorter than that for cortisol and most likely was the prednisolone metabolite tetrahydroprednisolone, an isomer of cortisol (4). The interference was not present in the m/z 36397 transition. As described previously, when an interference was detected for the m/z 363121 transition, quantification was performed based on the m/z 36397 transition to account for the interference. Compounds tested for potential interference with the method included prednisolone, prednisone, dexamethasone, desoxycorticosterone, fludrocortisone, 11-desoxycortisol, and the lipid-lowering drug fenofibrate (Tricor®). Two of the tested compounds showed potential for interference. Prednisolone produced a weak signal at a slightly different retention time. At the evaluated concentration of 5000 µg/L, prednisolone produced a signal equivalent to 20 µg/L cortisol, representing a 0.4% interference. Fenofibrate generated parent ions at m/z 361 and 363 in an abundance ratio of 3:1, a distinctive pattern related to the presence of chlorine in the chemical composition. In MS/MS mode, fenofibrate produced a m/z 363121 transition that interfered with the quantitative transition of cortisol, but it did not produce the m/z 36397 transition. Thus, switching to the secondary transition for quantification eliminated interference from this drug. In addition, the elution time of the drug was 30 s longer than that of cortisol.
Taylor et al. (1) noted isotopic exchange between the deuterated IS and hydrogen-containing vapors in an APCI ion source. Unlike the observations of Taylor et al., we have not observed isotopic exchange with our APCI interface. The reason for this difference is unclear, but it may be related to ion source conditions. The good agreement obtained between our method and the comparison LC-MS/MS method suggests that isotopic exchange did not affect the proposed method.
Apparently healthy adult volunteers (25 males and 25 females; age range, 19–53 years) collected 24-h urine samples without preservative. The volunteers were asked to keep samples under refrigeration during the collection. Statistical results for this study are presented in Table 1
.
|
To evaluate agreement between the established reference interval with UFC values in the population, we evaluated results for 2089 random 24-h urine specimens analyzed with the proposed method (Fig. 1
). The mean (SD) value for log-transformed UFC excretion in 24 h was 1.26 (0.24) log µg/24 h (minimum and maximum, 0.3 and 3.84 log µg/24 h, respectively), and skewness and kurtosis for the distribution were 1.51 and 5.57, respectively.
|
The stability of cortisol in urine was evaluated in the presence of acetic (15 mmol/L), boric (15 mmol/L), and hydrochloric (30 mmol/L) acid. Two samples without added acid were stored and analyzed under the same conditions as the samples stored with the acids. Samples were stored at room temperature, 4 °C, and -20 °C and analyzed every 4–7 days during 1 month of storage. Cortisol concentrations in samples stored with the acids were higher by 30% than in samples stored without acid, possibly as a result of partial hydrolysis of sulfate and glucuronide conjugates.
In conclusion, the rapid LC-MS/MS method for UFC analysis appears to be free from interference and agrees closely with a HPLC-MS/MS method that uses sample extraction. The method has been demonstrated reliable in a high-volume clinical laboratory environment.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  L. Wood, D. H Ducroq, H. L Fraser, S. Gillingwater, C. Evans, A. J Pickett, D. W Rees, R. John, and A. Turkes Measurement of urinary free cortisol by tandem mass spectrometry and comparison with results obtained by gas chromatography-mass spectrometry and two commercial immunoassays Ann Clin Biochem, July 1, 2008; 45(4): 380 - 388. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Kushnir, A. L. Rockwood, W. L. Roberts, E. G. Pattison, W. E. Owen, A. M. Bunker, and A. W. Meikle Development and Performance Evaluation of a Tandem Mass Spectrometry Assay for 4 Adrenal Steroids Clin. Chem., August 1, 2006; 52(8): 1559 - 1567. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. W. Meikle, J. Findling, M. M. Kushnir, A. L. Rockwood, G. J. Nelson, and A. H. Terry Pseudo-Cushing Syndrome Caused by Fenofibrate Interference with Urinary Cortisol Assayed by High-Performance Liquid Chromatography J. Clin. Endocrinol. Metab., August 1, 2003; 88(8): 3521 - 3524. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
