Original. Yosuke Sasaki 1), 2), Takuyuki Katabami 1), Shiko Asai 1), Hisashi Fukuda 1) and Yasushi Tanaka 2)

Transcription

1 2017, 64 (9), Original In the overnight dexamethasone suppression test, 1.0 mg loading is superior to 0.5 mg loading for diagnosing subclinical adrenal Cushing s syndrome based on plasma dexamethasone levels determined using liquid chromatography-tandem mass spectrometry Yosuke Sasaki 1), 2), Takuyuki Katabami 1), Shiko Asai 1), Hisashi Fukuda 1) and Yasushi Tanaka 2) 1) Division of Metabolism and Endocrinology, Department of Internal Medicine, St. Marianna University School of Medicine Yokohama City Seibu Hospital, Yokohama, Japan 2) Division of Metabolism and Endocrinology, Department of Internal Medicine, St. Marianna University School of Medicine, Kawasaki, Japan Abstract. The low-dose dexamethasone suppression test (DST) is one of the commonly used initial tests for endogenous Cushing s syndrome (CS). However, there are two loading dose regimens (0.5-mg and 1-mg), which may cause some confusion in daily practice in Japan; furthermore, there are no reports regarding whether 0.5-mg DST is a better loading dose for detecting adrenal subclinical CS (SCS) based on the plasma dexamethasone (DEX) levels. Therefore, the aims of this study were (a) to develop a novel assay to measure DEX by using liquid chromatography tandem-mass spectrometry (LC-MS/MS) and (b) to compare between the 0.5-mg and 1-mg DST for SCS diagnosis based on the DEX levels. The study retrospectively analyzed 52 consecutive subjects hospitalized for diagnosis of adrenal incidentaloma but who did not exhibit an overt CS phenotype; eight (15.4%) patients were affected with adrenal SCS. Inter-individual variability of DEX levels after the DST was high, but intra-individual variability was low. DEX levels after 1-mg loading in each patient was around two times higher than those after 0.5-mg loading (ρ = and p < 0.001). There were 45 (86.5%) and 17 (32.7%) subjects with DEX levels 2.2 ng/ml after the 0.5-mg and 1-mg DST, respectively (p < 0.001). Twenty-eight (93.3%) of 30 subjects and four (21.1%) of 19 subjects with detectable ACTH levels after the 0.5-mg and 1.0-mg DST, respectively, did not exhibit DEX levels >2.2 ng/ml. These results clearly indicate that the 1-mg DST is superior to 0.5-mg loading for the diagnosis of adrenal SCS. Key words: Dexamethasone suppression test, Plasma dexamethasone concentration, Liquid chromatography tandem-mass spectrometry, Adrenal subclinical Cushing s syndrome SUBCLINICAL CUSHING S SYNDROME (SCS) is characterized by subtle autonomous cortisol secretion from an adrenal incidentaloma (AI) without the phenotype of overt Cushing s syndrome (CS). This low-grade cortisol excess may worsen cardiovascular (CV) risk factors, increase CV events, and result in a higher mortality rate [1]. However, there is no consensus concerning establishment of a biochemical diagnosis for SCS [2]. The low-dose dexamethasone Submitted Feb. 20, 2017; Accepted May 4, 2017 as EJ Released online in J-STAGE as advance publication Jun. 21, 2017 Correspondence to: Takuyuki Katabami, M.D., Ph.D., Division of Metabolism and Endocrinology, Department of Internal Medicine, St. Marianna University School of Medicine Yokohama City Seibu Hospital, , Yasashicho, Asahi-ku, Yokohama, Kanagawa , Japan. t2kataba@marianna-u.ac.jp The Japan Endocrine Society suppression test (DST), originally described by Liddle in 1960 [3], is one of the commonly used initial tests for diagnosing endogenous CS. Administration of the potent synthetic glucocorticoid dexamethasone (DEX) suppresses the hypothalamic-pituitary-adrenal (HPA) axis in normal individuals, but SCS patients are at least partially resistant to this negative feedback. The simpler overnight low-dose DST was advocated by Nugent in 1965 [4]; doses ranging between 0.5 and 2 mg have subsequently been proposed for the DST, and various diagnostic cut-offs have been applied. Because of its greater simplicity, lower cost, and high sensitivity, recent published guidelines agree with the use of the 1-mg overnight DST for hypercortisolism screening [5].

2 834 Sasaki et al. According to diagnostic criteria for adrenal SCS set forth by the Intractable Disease Research Grant of the Ministry of Health, Labour, and Welfare (MHLW), Japan [6], cortisol levels >3 µg/dl following the 1-mg DST are used to diagnose adrenal SCS. Clinical practice guidelines from the Endocrine Society recommend use of the test, with a cortisol cut-off of <1.8 µg/dl, which is lower than previous criteria, to exclude autonomous cortisol secretion from AI [5]. In 2009, Oki et al. [7] described that a 0.5-mg DST with a serum cortisol cut-off level of 3 µg/dl may be more sensitive and specific for the diagnosis of ACTHdependent CS compared to the 1-mg DST, with a cut-off of 1.8 µg/dl. Accordingly, the two loading dose regimens may cause some confusion in daily practice, at least for general physicians. Furthermore, to avoid false positive and negative results, some experts have advocated simultaneous measurement of both cortisol and DEX for DST to ensure adequate plasma DEX levels (2.2 ng/ml) [8]. Although many studies report that DEX dose concentrations increase plasma DEX levels in a dose-dependent manner [9, 10], Oki et al. [7] did not report any data related to this. Due to cross-reactivity with possible metabolites of DEX, radioimmunoassays (RIAs) and enzyme-linked immunosorbent assays (ELISAs) require a highly selective antibody for accurate quantification of DEX, which is a serious drawback of these methods [11]. In contrast, structurally based assays such as liquid chromatography-tandem mass spectrometry (LC-MS/MS) do not present this problem [12]. Currently, there is no information regarding whether 0.5 mg is a better loading dose for detecting adrenal SCS based on plasma DEX levels. Therefore, the aim of this study was: (a) to develop a novel assay for DEX by using LC-MS/MS and (b) to compare between the 0.5-mg and 1-mg DST for SCS based on plasma levels of DEX in Japanese subjects with AI. Materials and Methods Subjects We retrospectively collected 52 consecutive subjects hospitalized for diagnosis of AI without any specific clinical features of CS between October 2013 and August 2015 at St. Marianna University School of Medicine Yokohama City Seibu Hospital (Yokohama, Japan). Patients taking steroids, oral estrogen, or antiepileptic agents were excluded, because these medications can influence the HPA axis [13]. Those with a history of alcohol abuse and psychiatric disorders were also excluded from this study. Hypertension was defined as the need for drug therapy, systolic blood pressure 140 mmhg, or diastolic pressure 90 mmhg. Diabetes mellitus (DM) was diagnosed as a need for antidiabetic drug therapy or a diabetic pattern on a 75-g oral glucose tolerance test. Hyperlipidemia was defined as the need for antilipidemic drug therapy, low-density lipoprotein cholesterol levels 140 mg/dl as calculated using Friedewald s formula [14], or triglyceride levels 150 mg/dl [15]. Patients with body mass index (BMI) 25 kg/m 2 were defined as obese. This study was approved by the Human Ethical Committee of St. Marianna University School of Medicine (No. 3324) and was performed according to the principles of the Helsinki Declaration. Dexamethasone suppression test The standard overnight 0.5-mg and 1-mg DST was performed in all subjects. Serum cortisol and plasma ACTH were measured at 0800 h after an overnight fast, and 0.5 mg DEX was administered at 2300 h on the same day. The next morning at 0800 h, serum cortisol, plasma ACTH, and plasma DEX measurements were obtained, and 1 mg DEX was administered at 2300 h. On day 3, a blood sample was collected again at 0800 h to complete the study. Nurses in our institute confirmed ingestion of DEX. According to Endocrine Society guidelines, false-positive results for DST were defined as 1) plasma DEX levels 2.2 ng/ml and 2) serum cortisol levels after DST above the cut-off value. False negative results were defined as 1) plasma DEX levels >2.2 ng/ml and 2) normal results for all endocrinological tests related to HPA axis function other than the DST. In the preliminary examination, we assayed plasma DEX levels at 24 h after 0.5-mg DEX administration in several patients with chronic adrenal insufficiency or non-functioning AI. Glucocorticoids were not detected in any sample, indicating that 0.5-mg DEX dose on day 2 is less likely to affect its plasma levels on day 3. Diagnostic method for adrenal SCS A diagnosis of SCS was defined as the presence of an AI, absence of specific clinical features of OCS, normal morning serum cortisol levels on more than two different evaluations, the presence of autonomous cortisol secretion confirmed by overnight 1-mg

3 1-mg DST is superior to 0.5-mg DST 835 DST ( 3.0 μg/dl serum cortisol following the DST), and at least one endocrine test including low plasma ACTH levels in the early morning (ACTH <10 pg/ml at 0800 h) or poor ACTH response to CRH (i.e., peak ACTH <30 pg/ml or less than 1.5 times greater than baseline), no diurnal changes in serum cortisol levels (serum cortisol levels 5.0 μg/dl at 2300 h), and low serum dehydroepiandrosterone sulfate (DHEA-S) levels (lower than age- and sex-matched reference levels). This diagnostic method followed the criteria set forth by the MHLW, Japan [6]. The Endocrine Society guidelines [5] and a recent report from Japan [16] proposed an alternative cut-off of serum cortisol >1.8 μg/ dl after the 1-mg DST for the diagnosis of adrenal SCS. Therefore, we also used this value for determination of DST results. Hormone assays Blood samples were immediately centrifuged at 4 C and then stored at 80 C until assayed. DHEA-S RIA kits (Mitsubishi Chemical Co., Tokyo, Japan), ACTH immunoradiometric assay kits (Mitsubishi Chemical Co.), and serum cortisol chemiluminescence enzyme immunoassay kits (Beckman Coulter, CA, USA) were used for hormone measurements. Measurements of other parameters To evaluate hepatorenal function, we also measured aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyltranspeptidase (γ-gtp), creatinine, and blood urea nitrogen. A sex-matched reference was used for the determination of γ-gtp levels (women: 9 36 IU/L and men: IU/L). LC-MS/MS DEX was obtained from Sigma-Aldrich (St. Louis, MO, USA). The deuterated internal standard, 4,6a,21, 21-d4-DEX (DEX-d 4 ) was obtained from C/D/N isotopes (Quebec, Canada). LC-MS grade solvents (methanol and water) and other reagents were purchased from Wako Pure Chemical Industries (Osaka, Japan). The sample was prepared as follows to measure dexamethasone. As an internal standard, DEX-d 4 dissolved in methanol was added to sample tubes, and methanol was evaporated to dryness. Human plasma (0.1 ml) was added to the sample tubes containing the internal standard. DEX was then extracted with 4 ml t-butyl methyl ether. After the organic layer was evaporated to dryness, the residue was dissolved in 0.5 ml methanol and diluted with 1 ml distilled water. The sample was applied to an Oasis MAX cartridge (60 mg, 3 ml, Waters Corporation, Milford, MA, USA), which had been successively conditioned with 3 ml of methanol and 3 ml of distilled water. The cartridge was successively washed with 1 ml distilled water, 1 ml methanol/distilled water/acetic acid (50:50:1, v/v/v), and 1 ml 1% pyridine. DEX was then eluted with 1 ml of methanol/distilled water/pyridine (80:20:1, v/v/v). After the solvent was evaporated, the residue was dissolved in 0.1 ml methanol/distilled water (2:3, v/v), and 10 µl of the solution was subjected to analysis in the LC-MS/MS instrument as described below. Each analytical standard (50 µl) of DEX (0, 50, 100, 500, 1,000, 5,000, 20,000 pg/μl) and internal standard (50 μl; 2,000 pg/50 μl) were added and evaporated. Distilled water/acetic acid (1 ml; 100:1, v/v) was then added to each solution to determine the calibration curves for DEX. As quality control (QC) samples, 50 μl of analytical standards of DEX (low QC: 75 pg/50 μl, middle QC: 2,000 pg/50 μl, high QC: 16,000 pg/50 μl) and 50 μl of internal standard (2,000 pg/50 μl) were added and evaporated. Distilled water/acetic acid (1 ml; 100:1, v/v) was then added to each solution. As a plasma matrix sample, 50 µl of an internal standard (2,000 pg/50 μl) was added and evaporated. The collected blood plasma (0.1 ml) and distilled water/acetic acid (0.9 ml; 100:1, v/v) were then added to the solution. DEX was analyzed using a technique similar to that in a previous report [17] using an API-4000 triple stage quadrupole mass spectrometer equipped with a positive electrospray ionization source (AB SCIEX, Framingham, MA, USA) and Nexera UHPLC systems (Shimadzu Co. Ltd., Kyoto, Japan). Regarding LC-MS/ MS analysis of DEX, 10 μl of the reconstituted samples were applied to a Kinetex C18 column ( mm, 1.7 μm, Shimadzu Co. Ltd., Kyoto, Japan). The mobile phase consisted of acetonitrile (Solvent A), and 0.1% formic acid (Solvent B) was used with a gradient elution of A:B = 25:75 (0 1.5 min), 35:65 (1.5 4 min), 100:0 (4 4.9 min), and 25:75 ( min) at a flow rate of 0.5 ml/min. The column temperature was maintained at 50 C. The column eluent was subjected to electron spray ionization (positive mode, ES+). DEX was determined using selective reaction monitoring of the transitions m/z , and were selected for DEX and DEX-d 4, respectively.

4 836 Sasaki et al. Statistical analysis Results are presented as the median (interquartile range [IQR]). Statistical analysis was performed using the Mann-Whitney U or Fisher s exact tests. Correlations among plasma DEX after DST, anthropometric measurements, and laboratory data were evaluated using Spearman s rank correlation analysis. The clinical factors were predicted plasma DEX levels by the univariate regression analysis. After identifying factors that influenced the plasma DEX levels by univariate analysis, multiple linear regression analysis was performed with the factors showing p <0.05 to find independent determinants of the plasma DEX levels. When data were skewed, logarithmic transformation was performed before multiple analysis. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) and graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria), a modified version of R commander that includes statistical functions frequently used in biostatistics. Statistical significance was defined as a p value < Results Validation of LC-MS/MS analysis As an additional recovery test, three concentrations of DEX (0, 500, and 1,000 pg) were spiked into human plasma samples to obtain samples with three different DEX levels. Subsequently, the spiked samples were analyzed in parallel with the calibration curve samples and QC samples by LC-MS/MS. The QC samples (three of each level, for a total of six samples) were analyzed before and after the human samples. Measured values were only adopted if the accuracy of measuring 4 QC samples was in the range of % and the two QC samples outside this reference range were not both samples with the same concentration. The acceptable range of the recovery rate was defined as %. In this test, the accuracy of all QC samples was in the range of %. The addition recovery rate was %. As a dilution test, three doses (0.01, 0.1, or 0.2 ml) of human plasma samples were spiked with DEX and analyzed in parallel with the calibration curve samples and QC samples using LC-MS/MS. The QC samples (three of each dilution level, for a total of six samples) were analyzed before and after the human plasma. Measured values were only adopted if the accuracy of measuring 4 QC samples was in the range of % and the two QC samples outside this reference range were not both samples with the same concentration. The acceptable error was defined as being within ±20% of the error calculated with an analytical standard of 0.2 ml. In the dilution test, the accuracy of measuring all QC samples was in the range of %, and the dilution level error was %. As a test of intra-assay reproducibility (betweenrun), five human plasma samples were spiked with DEX (50 pg) and were analyzed in parallel with the samples for the calibration curve and the QC samples by LC-MS/MS. The specified accuracy of measurement at 50 pg, the lower limit quantification (LLQ), was % [18], and the specified precision (the precisions below were shown coefficient validation) for human plasma samples was ±15%. The QC samples (three of each level, for a total of six samples) were analyzed before and after the human plasma. Measured values were only adopted if the accuracy of measuring 4 QC samples was in the range of % and the two QC samples outside this reference range were not both samples with the same concentration. In this test, the accuracy of measuring all QC samples was in the range of %, and the precision was %. Regarding inter-assay reproducibility (between-day), the accuracy and precision were % and %, respectively. The LLQ and quantitation range of DEX were 0.5 ng/ml and ng/ml, respectively, using 0.1 ml of plasma or 0.25 ng/ml and ng/ml, respectively, using 0.2 ml of plasma. Using 1 ml of plasma, the LLQ of this assay was 0.74 ng/ml, and the quantitation range was ng/ml. Patient characteristics Clinical and laboratory characteristics of the subjects at baseline are shown in Table 1. The median (IQR) age of 52 subjects was 65 (53, 73) years. The prevalence of obese, impaired glucose tolerance, hypertension, and dyslipidemia was 48.1%, 59.6%, 71.2%, and 46.2%, respectively. The median maximum diameter of adrenal tumor was 17.5 (12.8, 25.0) mm. Six of the 52 (11.5%) subjects were taking cytochrome P-450 (CYP) 3A4 inhibitors [19], such as clarithromycin (n = 1), cimetidine (n = 4), and diltiazem (n = 1), which could increase plasma DEX levels [5], while none of them were taking CYP3A4 inducers such as antiepileptic agents or rifampin [20]. No subjects were infected with hepatitis B or C virus or

5 1-mg DST is superior to 0.5-mg DST 837 had chronic hepatitis or liver cirrhosis. According to the MHLW criteria in Japan, 30 of 52 (57.7%) subjects were finally diagnosed as non-functional adrenal tumor. Seven subjects (13.5%) were diagnosed as SCS without primary aldosteronism (PA), one subject (1.9%) was diagnosed as SCS with PA, and 14 (26.9%) were diagnosed with PA. Thus, eight of 52 (15.4%) patients were affected with SCS. Plasma dexamethasone levels after the 0.5-mg or 1-mg DST Box-and-whisker plots of the morning plasma DEX levels after the 0.5- and 1-mg DST are shown in Fig. 1A. No patient had undetectable circulating plasma DEX levels, showing that all subjects had taken the drug. Irrespective of administration dose, plasma DEX values varied considerably, ranging from 0.2 to Table 1 Subject characteristics Total (n = 52) Female, n (%) 22 (42.3) Age, years 65 (53, 73) Height, cm 162 (157, 168) Body weight, kg 61 (56, 75) Body mass index, kg/m (21.9, 27.2) Body surface area, m (1.55, 1.82) Obesity, n (%) 25 (48.1) Impaired glucose tolerance, n (%) 31 (59.6) Hypertension, n (%) 37 (71.2) Dyslipidemia, n (%) 24 (46.2) Fasting plasma glucose, mg/dl 105 (95, 136) Glycoalbumin, % 20.9 (18.1, 26.8) HbA1c, % 6.2 (5.7, 7.7) AST, IU/L 21 (16, 25) ALT, IU/L 18 (13, 27) γ-gtp, IU/L 26 (17, 41) Total cholesterol, mg/dl 183 (153, 205) LDL-cholesterol, mg/dl 111 (87, 133) HDL-cholesterol, mg/dl 39 (35, 47) Triglycerides, mg/dl 104 (80, 145) Serum Cr, mg/dl 0.77 (0.63, 0.93) egfr, ml/min/1.73m (61.9, 81.6) Blood urea nitrogen, mg/dl 14 (11.3, 17.7) Tumor size, mm 17.5 (12.8, 25.0) 0800 h ACTH, pg/ml 10.9 (5.3, 22.3) 2400 h ACTH, pg/ml 6.2 (2.7, 9.6) 0800 h Cortisol, μg/dl 9.2 (7.9, 11.9) 2400 h Cortisol, μg/dl 5.2 (3.8, 6.4) Urinary free cortisol, μg/24 h 37.6 (26.2, 51.9) DHEA-S, μg/dl 59 (38.3, 113) NF 30 (57.7%), Final diagnosis, n (%) PA 14 (26.9%), SCS 7 (13.5%), PA+SCS 1 (1.9%) Data are expressed as the median and interquartile range (25 to 75%) for continuous variables or as percentages. γ-gtp, γ-glutamyltranspeptidase; HbA1c, glycated hemoglobin; AST, aspartate aminotransferase; ALT, alanine aminotransferase; LDL, low-density lipoprotein; HDL, high-density lipoprotein; Cr, creatinine; egfr, estimated glomerular filtration rate; ACTH, adrenocorticotropic hormone; DHEA-S, dehydroepiandrosterone sulfate; NF, non-functioning adrenal tumor; PA, primary aldosteronism; SCS, subclinical Cushing s syndrome. Fig. 1 Correlation of individual plasma DEX levels after the 0.5-mg DST and 1-mg DST Fig. 1A shows box-and-whisker plots of plasma DEX levels after the 0.5-mg DST and 1-mg DST. Plasma DEX values varied considerably. The lower (Q1) and upper (Q3) quartile, representing observations outside the 4 96 percentile range. The red solid lines represent the median of plasma DEX values. Horizontal broken lines (blue) represent the value of 2.2 ng/ml. However, as shown in Fig. 1B, the individual plasma DEX levels were significantly correlated (ρ = 0.853, p < 0.001), with a slope of and intercept of DEX, dexamethasone; DST, dexamethasone suppression test.

6 838 Sasaki et al. 5.1 ng/ml following the 0.5-mg DST and from 0.5 to 11.4 ng/ml following the 1-mg DST. In contrast, as shown in Fig. 1B, a highly significant statistical correlation was observed in plasma DEX levels between 0.5-mg and 1-mg after loading. Spearman s ρ and corresponding p value for these correlations were ρ = and p < 0.001, and slope and intercept were observed at and 0.072, respectively. Thus, these data clearly showed high inter-individual variability but low intra-individual variability of DEX levels after DST; furthermore, plasma levels after 1-mg loading in each patient were around two times higher than levels after 0.5-mg loading. Plasma dexamethasone and cortisol levels after the 0.5-mg or 1-mg DST Fig. 2 displays serum cortisol and plasma DEX levels after DST. The number of subjects with DEX levels 2.2 ng/ml after the 0.5-mg and 1-mg DST was 45 (86.5%) and 17 (32.7%), respectively (p < 0.001). Among eight subjects with a diagnosis of SCS, two subjects (25.0%) had plasma DEX levels >2.2 ng/ml and cortisol levels 3.0 μg/dl after DEX administration at a dose of 0.5-mg and the same two subjects (25.0%) showed DEX levels >2.2 ng/ml and cortisol levels >1.8 μg/dl with the same DEX dose (Fig. 2A). In contrast, four subjects (50.0%) had plasma DEX levels >2.2 ng/ml and cortisol levels 3.0 μg/ dl after a DEX dose of 1-mg, and the same four subjects (50.0%) showed DEX levels >2.2 ng/ml and cortisol levels >1.8 μg/dl (Fig. 2B). However, these four SCS subjects showed positive findings in tests of HPA axis function other than the DST, suggesting that low plasma DEX levels in the DST do not necessarily exclude SCS (data not shown). Plasma dexamethasone and ACTH levels after the 0.5-mg or 1-mg DST Detectable ACTH levels after the 0.5-mg and 1-mg DST were present in 30 (57.7%) and 19 (36.5%) subjects, respectively. Twenty-eight (93.3%) of the subjects with detectable ACTH levels after the 0.5-mg DST did not exceed a DEX level of 2.2 ng/ml; in contrast, only four (21.1%) of the subjects did so after the 1-mg DST. Fig. 2 Correlation between individual DEX and cortisol levels after DST Both panels show the individual distribution of DEX and cortisol levels after the 0.5-mg and 1-mg DST. Among eight subjects with a diagnosis of subclinical Cushing s syndrome (SCS), two subjects (25.0%) had plasma DEX levels >2.2 ng/ml and cortisol levels 3.0 μg/dl after DEX administration at a dose of 0.5 mg and the same two subjects (25.0%) showed DEX levels >2.2 ng/ml and cortisol levels >1.8 μg/dl with the same DEX dose (Fig. 2A). In contrast, four subjects (50.0%) had plasma DEX levels >2.2 ng/ml and cortisol levels 3.0 μg/dl after a DEX dose of 1 mg, and the same four subjects (50.0%) showed DEX levels >2.2 ng/ml and cortisol levels >1.8 μg/dl (Fig. 2B). Vertical broken lines (black) represent the value of 2.2 ng/ ml, upper horizontal broken lines (red) indicate the cut-off value of 3 μg/dl, lower horizontal broken lines (blue) indicate the cut-off value of 1.8 μg/dl, and red closed circles represent patients diagnosed as having SCS. DEX, dexamethasone; DST, dexamethasone suppression test.

7 1-mg DST is superior to 0.5-mg DST 839 Assessment of the relationship between plasma dexamethasone levels and clinical parameters We next performed univariate and multivariate analyses to identify clinical parameters related to plasma DEX levels after the 0.5-mg or 1-mg DST (Table 2). In the Spearman s rank correlation analysis, there was a weak correlation between DEX levels after the 0.5-mg DST and body mass index (BMI) (ρ = 0.273, p = 0.049). There were also weak correlations between DEX levels after the 1-mg DST and body weight (ρ = 0.285, p = 0.041), BMI (ρ = 0.328, p = 0.018), γ-gtp (ρ = 0.339, p = 0.014), and glycated hemoglobin (ρ = 0.290, p = 0.037). The plasma DEX levels in the subject taking CYP3A4 inhibitors tended to be slightly higher, but not to a statistically significant level (data not shown). Furthermore, multiple linear analysis revealed that there were no independent variables for 0.5-mg DST, while γ-gtp was selected as a significant independent variable associated with the plasma DEX levels after the 1-mg DST. However, the association between γ-gtp and DEX levels was lost after excluding subjects with γ-gtp levels above the sex-matched reference range. Discussion There were two main features of the present study. First, we reported the development of an accurate and precise LC-MS/MS assay for measuring plasma DEX levels. Second, the 1-mg DST was superior to 0.5-mg loading for the diagnosis of adrenal SCS based on the plasma DEX levels measured using the LC-MS/ MS assay. As has been indicated, because inadequate plasma DEX levels may cause false-positive results on the DST, quantitative determination of synthetic glucocorticoid is indispensable verification for comparing the accuracy between DST procedures. LC-MS/MS offers superior analytical specificity compared with conventional immunoassays. Moreover, an antibody to DEX is not currently available in Japan. Therefore, we developed a LC-MS/MS method to quantify DEX. DEX levels displayed acceptable standard deviations ( %) and a low LLQ (0.74 ng/ml). Results of the additional recovery test and the reproducibility test were also satisfactory, showing that the assay system we developed is reliable enough to undergo further investigation. Table 2 Results of univariate (upper) and multiple linear (lower) regression analysis of factors affecting plasma DEX levels after the 0.5-mg or 1-mg DST 0.5-mg DST 1-mg DST ρ p value ρ p value Age, years Body weight, kg Body mass index, kg/m Height, cm egfr, ml/min/1.73m Serum Cr, mg/dl Blood urea nitrogen, mg/dl AST, IU/L ALT, IU/L γ-gtp, IU/L HbA1c, % mg DST 1-mg DST B β p value B β p value Body mass index, kg/m γ-gtp, IU/L HbA1c, % Upon multiple linear regression analysis, γ-gtp was a significant parameter for affecting plasma DEX levels after 1-mg loading. Because body mass index (BMI) shows an interaction with body weight and had a smaller p value on univariate analysis, BMI was used as an explanatory variable. Values of p <0.05 were considered statistically significant. B, partial regression coefficient; β, standardized partial regression coefficient; DEX, dexamethasone; DST, dexamethasone suppression test; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-gtp, γ-glutamyltranspeptidase; Cr, creatinine; egfr, estimated glomerular filtration rate.

8 840 Sasaki et al. No study has directly compared plasma DEX levels measured by using LC-MS/MS and a conventional method. Nonetheless, plasma DEX levels determined using LC-MS/MS at 0800 h following 1-mg DST were similar to those determined using an RIA method, with plasma levels in most subjects within the range of 1.0 ng/ml to 10 ng/ml [8, 21-23]. The present study shows the superiority of the 1-mg DST to 0.5-mg DST; this conclusion is supported by the following reasons: 1) plasma DEX levels after the 1-mg DST in each patient were consistently higher than those after 0.5-mg loading, 2) use of 1-mg DEX can more frequently achieve the recommended value by Endocrine Society guidelines, and 3) ACTH secretion was more strongly suppressed by 1-mg DEX than by 0.5-mg DEX. Nevertheless, a lower loading dose and simplification of DST is beneficial and desirable from a safety standpoint, and the use of recent less-stringent criteria for the use of cortisol in testing is expected to prevent underestimation of subtle hypercortisolism, including SCS. However, these alterations may increase the risk for false-positive DST results; additionally, using a recent procedure, the diagnostic specificity in some studies of CS or pseudo-cs was lower than that previously reported [24]. Accordingly, it is important to develop procedures for simultaneous determination of post-test DEX and cortisol levels. On the 1-mg DST, the frequency of false-positive results remained approximately 10%, but reached 20 40% on the 0.5-mg DST. Thus, there is no rational basis for changing the dose of DEX from 1 mg to 0.5 mg. DEX levels after a low-dose DST in the current study did not attain a predictably effective value (2.2 ng/ml) in a higher proportion of patients than previously reported [21]. One possible explanation for the unexpected results is that the Endocrine Society recommendations regarding DEX levels may be inadequate. The recommendation was based on results indicating that when the plasma DEX levels following overnight low dose DST were <2.2 ng/ml, there was an overlap of the range of cortisol levels between patients with documented CS and those with suspected CS but who exhibited no progression of Cushingoid features on follow-up [8]. Furthermore, the report used the classic threshold value of 5 µg/dl for inadequate cortisol suppression on DST; hence, it is uncertain whether the results can also apply to the diagnosis of SCS. Accordingly, further research is needed to clarify the optimal cut-off plasma DEX value for evaluation of autonomous, albeit subtle, cortisol secretion from adrenal adenoma. There is considerable inter-individual variation in plasma DEX levels after DST [8]. This may be due to several factors that may affect oral DST; such factors include variable intestinal absorption and decreased or increased DEX metabolism due to other compounds. Moreover, age, sex, and adiposity might influence the outcome of DST, but results vary among studies [10, 25, 26]. From physical differences between Japanese and Western people, Oki et al. [7] suggested that in Japanese patients, 1-mg DST might suppress plasma cortisol levels too strongly. Using a receiver operator characteristic analysis, Oki et al. [7] concluded that 0.5-mg DST is a more sensitive and specific screening tool for diagnosis of ACTH-dependent CS compared to 1-mg DST. However, this prior study did not mention DEX levels. In the present study, body weight, body surface area, and BMI were not associated with plasma DEX levels on multiple regression analysis, indicating that physical differences do not provide a rationale for adjustment of DEX dose in patients with suspected adrenal SCS. In agreement with our results, most studies showing differences in body weight fail to account for the marked variability in post-oral administration DEX levels [23, 25]. O Sullivan et al. indicated no single pharmacokinetic factor was responsible for the lower DEX levels in depressed patients without adequate cortisol suppression by DST [26]. Moreover, Weiner et al. reported that healthy subjects with DEX 2.2 ng/ml were rare (5%), but all were >50 years old, indicating that DST may be invalid more often in older individuals [27]. Although the cause of differences among studies is not clear, these findings indicate that there may be an increased need for simultaneous measurement of cortisol and DEX in older individuals. Additionally, degradation of DEX by hepatic CYP3A4 also may be responsible for the variation in DEX levels; several drugs markedly accelerate glucocorticoid clearance, resulting in much lower levels than expected [5]. We tried to exclude patients taking CYP3A4 inhibitors as much as possible, but six patients wished to continue taking them. However, in this study plasma DEX levels in these subjects were not significantly higher than that in other subjects, suggesting that the agents may not have affected the present findings. This further suggests that medication use should be considered when administering DST.

9 1-mg DST is superior to 0.5-mg DST 841 Furthermore, Huizenga et al. [10] reported a positive correlation between AST, ALT, and γ-gtp levels with DEX levels after 1-mg DEX administration in men from a population-based cohort. Nonetheless, the relationship was no longer significant when excluding subjects with supra-normal levels of at least one of the three enzymes. Although we excluded patients with positive hepatitis B surface antigen and hepatitis C virus antibody, chronic hepatitis, liver cirrhosis, or alcohol abuse, we likewise found a significant association between γ-gtp and plasma DEX levels after the 1-mg DST. Interestingly, the association was also lost after excluding subjects with γ-gtp above the sex-matched reference range. These findings suggest that subjects without chronic alcoholic or viral liver disorder but with mild liver dysfunction may demonstrate false-positive results on a DST. In a previous study, nonalcoholic fatty liver disease (NAFLD) occurred in 20% of patients with CS, and glucocorticoid excess appeared to increase hepatic de novo lipogenesis [28]. Recently, Woolsey et al. [29] showed that CYP3A4 activity was significantly reduced in human NAFLD as well as in mouse and cell culture models of NAFLD. Thus, mild liver dysfunction resulting from NAFLD may be present in some SCS patients and may delay DEX metabolism via reduced CYP3A4 function [30]. Further investigation to determine the importance of this issue will be required. This study has several limitations. First, we did not determine whether the validity of the conventional cut-off value for plasma DEX levels (2.2 ng/ ml), as previously set forth by the RIA method [8], also applies to results obtained using the LC-MS/MS method. However, the purpose of this study was to compare the precision of the 1-mg and 0.5-mg DST based on plasma DEX levels or magnitude of ACTH suppression. Notwithstanding, even if the cut-off value changed, we believe that the superiority of the 1-mg DST to 0.5-mg loading still applies. Second, our patients did not undergo imaging studies or liver biopsy for the diagnosis of NAFLD, which may affect DEX metabolism via decreased CYP3A4 activity in patients with AI; as this has rarely been investigated, this is an important issue to be addressed in the future. Third, we tried to exclude subjects with diseases or those taking drugs that may potentially affect cortisol-binding globulin (CBG), but we did not measure CBG levels because an antibody against this protein is not available in Japan. Finally, the present study lacked a healthy control group, and the number of subjects was small. Therefore, further large-scale cohort studies are needed to confirm our results. In conclusion, we developed a precise LC-MS/ MS method for measuring DEX levels in plasma. Furthermore, we demonstrated that the 1-mg DST can achieve consistently higher plasma DEX levels and less frequent false-positive results compared to those with the 0.5-mg DST, showing that the 1-mg DST is superior to the 0.5-mg DST for the diagnosis of adrenal SCS. Acknowledgments This study was supported by a Health and Labor Science Research Grant, Research on Intractable Diseases, Research Committee on Disorder of Adrenal Hormones from the MHLW, Japan. The authors wish to thank Dr. Hideki Katakami, Department of Medicine, Kohnan Kakogawa Hospital for providing insightful comments and suggestions. Disclosure None of the authors have any potential conflicts of interest associated with this research. References 1. Di Dalmazi G, Pasquali R, Beuschlein F, Reincke M (2015) Subclinical hypercortisolism: a state, a syndrome, or a disease? Eur J Endocrinol 173: M Zografos GN, Perysinakis I, Vassilatou E (2014) Subclinical Cushing s syndrome: current concepts and trends. Hormones (Athens) 13: Liddle GW (1960) Tests of pituitary-adrenal suppressibility in the diagnosis of Cushing s syndrome. J Clin Endocrinol Metab 20: Nugent CA, Nichols T, Tyler FH (1965) Diagnosis of Cushing s syndrome: Single dose dexamethasone suppression test. Arch Intern Med 116: Nieman LK, Biller BM, Findling JW, Newell-Price J, Savage MO, et al. (2008) The diagnosis of Cushing s syndrome: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 93: Nawata H, Demura H, Suda T, Takayanagi R (1996) Adrenal preclinical Cushing s syndrome, Annual report

10 842 Sasaki et al. of the Ministry of Health and Welfare Disorder of Adrenal Hormones Research Committee, Japan 1995: (In Japanese). 7. Oki Y, Hashimoto K, Hirata Y, Iwasaki Y, Nigawara T, et al. (2009) Development and validation of a 0.5 mg dexamethasone suppression test as an initial screening test for the diagnosis of ACTH-dependent Cushing s syndrome. Endocr J 56: Meikle AW (1982) Dexamethasone suppression tests: usefulness of simultaneous measurement of plasma cortisol and dexamethasone. Clin Endocrinol (Oxf) 16: Loew D, Schuster O, Graul EH (1986) Dose-dependent pharmacokinetics of dexamethasone. Eur J Clin Pharmacol 30: Huizenga NA, Koper JW, de Lange P, Pols HA, Stolk RP, et al. (1998) Interperson variability but intraperson stability of baseline plasma cortisol concentrations, and its relation to feedback sensitivity of the hypothalamo-pituitary-adrenal axis to a low dose of dexamethasone in elderly individuals. J Clin Endocrinol Metab 83: Murakami M, Miyachi Y, Nanno M, Yoshimi T (1990) Radioimmunoassay for dexamethasone and its plasma levels after oral administration in patients with liver disease. Folia Endocrinologica Japonica 66: (In Japanese). 12. Djedovic NK, Rainbow SJ (2011) Detection of synthetic glucocorticoids by liquid chromatography-tandem mass spectrometry in patients being investigated for Cushing s syndrome. Ann Clin Biochem 48: Nickelsen T, Lissner W, Schöffling K (1989) The dexamethasone suppression test and long-term contraceptive treatment: measurement of ACTH or salivary cortisol does not improve the reliability of the test. Exp Clin Endocrinol 94: Friedewald WT, Levy RI, Fredrickson DS (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18: Arnaldi G, Scandali VM, Trementino L, Cardinaletti M, Appolloni G, et al. (2010) Pathophysiology of dyslipidemia in Cushing s Syndrome. Neuroendocrinology 92: Akehi Y, Kawate H, Murase K, Nagaishi R, Nomiyama T, et al. (2013) Proposed diagnostic criteria for subclinical Cushing s syndrome associated with adrenal incidentaloma. Endocr J 60: Wooding KM, Auchus RJ (2013) Mass spectrometry theory and application to adrenal diseases. Mol Cell Endocrinol 371: Methlie P, Hustad SS, Kellmann R, Almås B, Erichsen MM, et al. (2013) Multisteroid LC-MS/MS assay for glucocorticoids and androgens, and its application in Addison s disease. Endocr Connect 2: Kanazu T, Sato N, Kadono K, Touchi A, Takeda Y, et al. (2012) Investigation of drug-drug interaction via mechanism-based inhibition of cytochrome P450 3A by macrolides in dexamethasone-treated female rats. Biopharm Drug Dispos 33: Kyriazopoulou V, Vagenakis AG (1992) Abnormal overnight dexamethasone suppression test in subjects receiving rifampicin therapy. J Clin Endocrinol Metab 75: Hempen C, Elfering S, Mulder AH, van den Bergh FA, Maatman RG (2012) Dexamethasone suppression test: development of a method for simultaneous determination of cortisol and dexamethasone in human plasma by liquid chromatography/tandem mass spectrometry. Ann Clin Biochem 49: Ray JA, Kushnir MM, Rockwood AL, Meikle AW (2011) Analysis of cortisol, cortisone and dexamethasone in human serum using liquid chromatography tandem mass spectrometry and assessment of cortisol: Cortisone ratios in patients with impaired kidney function. Clin Chim Acta 412: Åsvold BO, Grill V, Thorstensen K, Bjørgaas MR (2012) Association between posttest dexamethasone and cortisol concentrations in the 1 mg overnight dexamethasone suppression test. Endocr Connect 1: Pecori Giraldi F, Pivonello R, Ambrogio AG, De Martino MC, De Martin M, et al. (2007) The dexamethasone-suppressed corticotropin-releasing hormone stimulation test and the desmopressin test to distinguish Cushing s syndrome from pseudo-cushing s states. Clin Endocrinol (Oxf) 66: Morris H, Carr V, Gilliland J, Hooper M (1986) Dexamethasone concentrations and the dexamethasone suppression test in psychiatric disorders. Br J Psychiatry 148: O Sullivan BT, Cutler DJ, Hunt GE, Walters C, Johnson GF, et al. (1997) Pharmacokinetics of dexamethasone and its relationship to dexamethasone suppression test outcome in depressed patients and healthy control subjects. Biol Psychiatry 41: Weiner MF (1989) Age and cortisol suppression by dexamethasone in normal subjects. J Psychiatr Res 23: Rockall AG, Sohaib SA, Evans D, Kaltsas G, Isidori AM, et al. (2003) Hepatic steatosis in Cushing s syndrome: a radiological assessment using computed tomography. Eur J Endocrinol 149: Woolsey SJ, Mansell SE, Kim RB, Tirona RG, Beaton MD (2015) CYP3A Activity and Expression in Nonalcoholic Fatty Liver Disease. Drug Metab Dispos 43: Tarantino G, Finelli C (2013) Pathogenesis of hepatic steatosis: The link between hypercortisolism and nonalcoholic fatty liver disease. World J Gastroenterol 19: