Hindawi International Journal of Endocrinology Volume 2019, Article ID 7251010, 8 pages https://doi.org/10.1155/2019/7251010 Research Article Salivary Cortisol Determination in ACTH Stimulation Test to Diagnose Adrenal Insufficiency in Patients with Liver Cirrhosis Lara Albert ,1 Joaqu-m Profitós,2 Jordi Sánchez-Delgado ,2,3 Ismael Capel ,1 José Miguel González-Clemente,1 David Sub-as,1 Albert Cano ,1 Eugenio Berlanga,4 Anna Espinal,5 Marta Hurtado,1 Roc-o Pareja,1 Mercedes Rigla ,1 Blai Dalmau,2 Mercedes Vergara,2,3 Mireia Miquel,2,3 Meritxell Casas,2 and Olga Giménez-Palop1 1 Endocrinology Department, Parc Taulı́ University Hospital, Institut d’Investigacio i Innovació Parc Taulı́ I3PT, Autonomous University of Barcelona, Sabadell, Spain 2 Liver and Gastrointestinal Diseases Department, Parc Tauli University Hospital, Institut d’Investigacio i Innovació Parc Taulı́ I3PT, Autonomous University of Barcelona, Sabadell, Spain 3 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Spain 4 Clinical Laboratory Department, Institut d’Investigació i Innovació Parc Taulı́ I3PT, Parc Taulı́ University Hospital, Autonomous University of Barcelona, Sabadell, Spain 5 Servei d’Estadı́stica Aplicada, Universitat Autònoma de Barcelona, Bellaterra, Spain Correspondence should be addressed to Jordi Sánchez-Delgado; jsanchezd@tauli.cat Received 14 March 2019; Revised 13 May 2019; Accepted 19 May 2019; Published 20 June 2019 Academic Editor: Giuseppe Reimondo Copyright © 2019 Lara Albert et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Purpose. The prevalence of adrenal insufficiency (AI) in patients with decompensated liver cirrhosis is unknown. Because these patients have lower levels of cortisol-binding carrier proteins, their total serum cortisol (TSC) correlates poorly with free serum cortisol (FC). Salivary cortisol (SaC) correlates better with FC. We aimed to establish SaC thresholds for AI for the 250 𝜇g intravenous ACTH test and to estimate the prevalence of AI in noncritically ill cirrhotic patients. Methods. We included 39 patients with decompensated cirrhosis, 39 patients with known AI, and 45 healthy volunteers. After subjects fasted ≥8 hours, serum and saliva samples were collected for determinations of TSC and SaC at baseline 0’(T0 ) and at 30-minute intervals after intravenous administration of 250 𝜇g ACTH [30’(T30 ), 60’(T60 ), and 90’(T90 )]. Results. Based on the findings in healthy subjects and patients with known AI, we defined AI in cirrhotic patients as SaC-T0 < 0.08 𝜇g/dL (2.2 nmol/L), SaC-T60 < 1.43 𝜇g/dl (39.5 nmol/L), or ΔSaC<1 𝜇g/dl (27.6 nmol/L). We compared AI determination in cirrhotic patients with the ACTH test using these SaC thresholds versus established TSC thresholds (TSC-T0 < 9 𝜇g/dl [248 nmol/L], TSC-T60 < 18 𝜇g/dl [497 nmol/L], or ΔTSC<9 𝜇g/dl [248 nmol/L]). SaC correlated well with TSC. The prevalence of AI in cirrhotic patients was higher when determined by TSC (48.7%) than by SaC (30.8%); however, this difference did not reach statistical significance. AI was associated with sex, cirrhosis etiology, and Child-Pugh classification. Conclusions. Measuring SaC was more accurate than TSC in the ACTH stimulation test. Measuring TSC overestimated the prevalence of AI in noncritically ill cirrhotic patients. 1. Introduction symptoms of acute and chronic AI makes the diagnosis difficult. Adrenal insufficiency (AI) is common in patients with liver The insulin tolerance test is considered the gold stan- disease; AI is present in both patients with severe cirrhosis dard for evaluating the hypothalamus-pituitary–adrenal axis. admitted to intensive care units and stable patients [1–3]. However, in clinical practice, the short ACTH stimulation Relative AI in patients with cirrhosis is a sum of primary test is more widely used because it is better tolerated and has (lack of steroid precursors, such as cholesterol) and secondary fewer contraindications. Both tests are based on the analysis (impairment of CRH-ACTH axis) AI. The lack of specific of serum cortisol, and the correlation between the two is well 2 International Journal of Endocrinology studied; most clinical guidelines support the use of the ACTH Table 1: Clinical characteristics of cirrhotic patients. test for diagnosing AI [4, 5]. Variable Data Serum cortisol is mostly bound to carrier proteins such as cortisol-binding globulin (CBG) and albumin [6]. Free Age, years∗ 58.5 ± 11.7 cortisol (FC), the biologically active unbound fraction, rep- Age at diagnosis of cirrhosis, years∗ 53.8 ± 10.8 resents about 5% to 10% of total serum cortisol (TSC) [7, Male sex, n (%) 34 (87.2) 8]. Various conditions affect protein synthesis. For instance, Cirrhosis etiology, n (%) cirrhosis, malnutrition, and critical illness reduce it, whereas Alcohol 30 (76.9) oral contraceptives and pregnancy increase it. Thus, TSC does Alcohol and HCV 5 (12.8) not accurately reflect FC, increasing the risk of misdiagnosis Alcohol and HBV 2 (5.1) [9–11]. Rauschecker et al. [12] recently demonstrated that Primary biliary cirrhosis 1 (2.6) measuring FC in response to ACTH stimulation is a good Autoimmune disorder 1 (2.6) alternative to TSC for diagnosing AI. However, FC analysis is Type 2 diabetes, n (%) 9 (23) time-consuming and expensive, hindering its use for routine laboratory testing. The FC fraction can be calculated using the Ascites/edema, n (%) 35 (89.7) Coolens’ equation, but the results are unsatisfactory [13]. An Encephalopathy, n (%) 9 (23.1) easier, less expensive approach is to determine FC indirectly Hypertension, n (%) 12 (30.8) by measuring salivary cortisol (SaC), a surrogate of plasma Child-Pugh score∗ 9±2 FC [14]. Class A/B/C, n (%) 6 (15.4)/18 (46.2)/15 (38.5) Late-night (23:00–24:00 h) SaC is widely used to detect MELD score∗ 17 ± 6 hypercortisolism when Cushing’s syndrome is suspected [15]. Albumin, g/L∗ 30.4 ± 6.4 Various authors have proposed using SaC instead of TSC after Prealbumin, g/L∗ 6.0 ± 2.4 ACTH stimulation tests [15–17], but limited data are available Cholesterol, mg/dL𝛿 109 (83-134) to validate this approach. HDL cholesterol, mg/dL𝛿 28 (13-39) We aimed to determine the reference values for SaC after stimulation with 250 𝜇g intravenous ACTH, to determine LDL cholesterol, mg/dL𝛿 68 (50-78) the diagnostic accuracy of these values for AI in noncritical Triglycerides, mg/dL𝛿 72 (56-105) patients with cirrhosis, and to estimate the prevalence of AI AST, IU/L𝛿 47 (34-61) in this population. ALT, IU/L𝛿 21 (15-31) GGT, IU/L𝛿 108 (39-225) 2. Materials and Methods Bilirubin, mg/dL𝛿 2.6 (1.6-5.8) Prothrombin time∗ 1.68 ± 0.91 Subjects were enrolled from April 2013 through October 2015. INR∗ 1.80 ± 0.98 We included (a) 45 healthy adults recruited from hospital ∗Mean ± SD; 𝛿 median (Q1-Q3). staff (17 men; mean age, 30 years; range, 22–49 years); none required medication within 1 month of testing, and all had normal liver, renal, and thyroid function; (b) 41 endocrinology patients with known AI diagnosed by insulin corticosteroid replacement therapy received their last dose at tolerance test or short 250𝜇g ACTH test (12 men; mean age, 9:00 a.m. the day before testing. 57 years; range, 24–86 years; 13 primary AI, 26 secondary The first sample was extracted between 8:30 a.m. and AI); and (c) 39 noncritical cirrhotic patients hospitalized for 9:30 a.m. after at least 8h fasting. To avoid stress-induced cirrhosis-related complications (34 men; mean age, 58 years; bias, baseline (T0 ) samples were obtained 30 minutes after range, 39–90 years). Cirrhosis was diagnosed through histo- catheterization of a superficial vein. Blood samples were logical or clinical, biological, and ultrasonographic findings. drawn from the catheter. Saliva samples were collected after Table 1 reports cirrhotic patients’ demographic and clinical patients chewed a cotton swab specially designed for corti- data. Reasons for hospitalization were ascites, gastrointestinal sol determination from saliva (Salivette, Sarstedt AG&Co; bleeding, infection without systemic inflammatory response Nümbrecht, Germany) for 1 to 3 minutes. After 250 𝜇g of syndrome, alcoholic hepatitis, acute kidney injury, and oth- synthetic ACTH (Synacthen, Alfasigma; Milan, Italy) was ers. administered intravenously, blood and saliva samples were Exclusion criteria were age <18 years; pregnancy; use of collected at minutes 30 (T30 ), 60 (T60 ), and 90 (T90 ). All glucocorticoids (except in AI patients) or oral contraceptives samples were processed immediately. < 6 months before inclusion; severe acute illness; mean SaC and TSC levels were determined by electrochemi- arterial pressure < 60 mmHg; blood in the mouth; adminis- luminescence assay (Roche Diagnostics GmbH; Mannheim, tration of albumin, fresh frozen plasma, or terlipressin before Germany) [lower limit of detection, 0.018 𝜇g/dL (0.50 inclusion; or absence of consent. nmol/L); coefficient of variation, 4.1%–4.9% at high levels and At inclusion, patients were examined, with special atten- 7.5%–11.5% at low levels]. Values of SaC or TSC <0.018 𝜇g/dL tion to the presence of blood in the mouth. Patients were (0.50 nmol/L) were excluded from the analyses. told not to brush their teeth, smoke, or drink anything but For the diagnosis of AI, we used the following established water during the 60 min before sampling. Patients on chronic cutoffs: TSC T0 < 9 𝜇g/dl (248 nmol/L), TSC T60 < 18 International Journal of Endocrinology 3 Table 2: Serum cortisol concentrations in healthy volunteers, adrenal insufficient patients, and cirrhotic patients in the 250 𝜇g ACTH test. Serum cortisol (𝜇g/dL) Mean Median SD Min Q1 Q3 Max Group T0 15.15 13.95 5.66 6.14 11.18 18.48 29.61 T30 26.45 27.34 4.64 18.74 23.15 29.63 41.26 Healthy T60 31.12 32.23 4.60 22.09 27.38 34.01 45.24 T90 34.22 34.55 4.72 26.29 30.62 37.68 47.68 T0 1.94 0.67 2.78 0.03 0.21 2.45 10.10 T30 3.87 2.23 4.61 0.03 0.68 5.07 17.53 Adrenal insufficiency T60 4.78 2.51 5.52 0.03 0.84 6.42 21.20 T90 5.22 2.54 6.14 0.03 0.84 7.09 23.30 T0 12.75 12.27 5.91 2.05 8.20 17.80 24.85 T30 21.52 22.35 6.79 11.02 16.35 26.33 40.20 Cirrhosis T60 26.23 26.73 7.93 13.83 18.59 32.96 42.76 T90 29.94 29.89 9.01 15.28 21.31 38.17 45.67 T0 , baseline; T30 , 30 min after administration; T60 , 60 min after administration; T90 , 90 min after administration. To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. Table 3: Change in serum cortisol between baseline and 60 minutes in healthy volunteers, adrenal insufficient patients, and cirrhotic patients in the 250 𝜇g ACTH test. Delta serum cortisol (𝜇g/dL) Mean Median Std Min Q1 Q3 Max Group Healthy 15.97 15.63 4.80 7.31 12.30 19.63 25.34 Adrenal insufficiency 2.83 1.73 3.14 -0.04 0.19 5.17 12.33 Cirrhosis 13.48 13.06 5.36 4.23 8.70 16.51 25.56 To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. 𝜇g/dl (497 nmol/L), or ΔTSC (increase between T0 and T60 ) 𝜇g/dL (441±132 nmol/L) in healthy subjects and 2.84±3.14 <9 𝜇g/dl (248 nmol/L) [4, 18–20]. Salivary cortisol cutoffs 𝜇g/dL (78±87 nmol/L) in AI patients (Table 3). were defined as the minimum SaC concentration observed In healthy subjects, mean SaC at T0 was 0.56 ± 0.31𝜇g/dL in healthy subjects at T0 and T60 and the minimum ΔSaC (15±9 nmol/L); the lowest value was 0.08 𝜇g/dL (2.2 nmol/L) value. We analyzed the correlation between TSC levels and (Table 4). The area under the receiver operating characteristic SaC levels. We used the SaC cutoffs and TSC cutoffs to assess curve for SaC-T0 was 0.8045. After ACTH stimulation, SaC the prevalence of AI in the cirrhotic group and compared the progressively increased in nearly all healthy subjects; the results obtained with the two methods. lower limit of SaC at T60 was 1.43 𝜇g/dL (39.5 nmol/L) We did a descriptive analysis of patients’ clinical char- (Table 4). The cutoff SaC at T60 > 1.43 𝜇g/dL (39.5 nmol/L) acteristics. We used descriptive statistics to summarize the classified all AI patients correctly. values of TSC and SaC at each timepoint and the differences In healthy subjects, mean ΔSaC was 1.79±0.59 𝜇g/dL between their values at baseline and 60 minutes (ΔTSC and (49.4±16.3 nmol/L); the lowest value was 1 𝜇g/dL (27.6 ΔSaC). Using these statistics, we defined three criteria for the nmol/L) (Table 5). We defined these concentrations (SaC- diagnosis of AI. The likelihood-ratio test was used to check T0 < 0.08 𝜇g/dL [2.2 nmol/L], SaC-T60 < 1.43 𝜇g/dL [39.5 the goodness of fit. nmol/L], ΔSaC<1 𝜇g/dL [27.6 nmol/L]) as cutoff values for the We used SAS 9.3 (SAS System, Cary, NC, USA, 2013) for diagnosis of AI. all analyses. In AI patients, mean SaC at T0 was 0.33±0.30 𝜇g/dL (9.1±8.3 nmol/L); the highest value was 1.53𝜇g/dL (42.2 3. Results nmol/L) (Table 4). After ACTH stimulation, SaC in AI 3.1. Healthy and AI Subjects. All healthy controls had TSC ≥18 patients mainly remained constant over time (Table 4). 𝜇g/dL (497 nmol/L) at T60 ; two AI subjects surpassed this The highest concentration of SaC at T60 in AI patients cutoff and were excluded from the analyses. Figure 1 shows was 0.90𝜇g/dL (24.8 nmol/L); therefore, the SaC-T60 cutoff the distribution of TSC and SaC after ACTH stimulation. As classified all AI patients correctly. Mean ΔSaC in AI patients expected, all TSC determinations were lower in AI subjects was -0.004±0.18 𝜇g/dL (-0.11±5.0 nmol/L); all had ΔSaC than in healthy subjects (Table 2). Mean ΔTSC was 15.97±4.80 values lower than 1 𝜇g/dL (27.6 nmol/L) (Table 5). 4 International Journal of Endocrinology Table 4: Salivary cortisol concentrations in healthy volunteers, adrenal insufficient patients, and cirrhotic patients in the 250 𝜇g ACTH test. Salivary cortisol (𝜇g/dL) Mean Median Std Min Q1 Q3 Max Group T0 0.56 0.52 0.31 0.08 0.37 0.68 1.37 T30 1.58 1.56 0.45 0.83 1.28 1.83 2.72 Healthy T60 2.35 2.19 0.63 1.43 1.89 2.78 4.24 T90 2.92 2.71 0.82 1.63 2.35 3.40 5.42 T0 0.33 0.26 0.30 0.05 0.13 0.36 1.53 T30 0.32 0.29 0.24 0.05 0.13 0.41 1.17 Adrenal insufficiency T60 0.32 0.27 0.24 0.05 0.12 0.42 0.90 T90 0.37 0.28 0.50 0.05 0.11 0.37 3.13 T0 0.72 0.69 0.42 0.14 0.35 0.99 1.88 T30 1.47 1.14 0.96 0.37 0.80 1.73 5.19 Cirrhosis T60 2.25 1.97 1.20 0.58 1.47 2.90 5.37 T90 3.12 2.67 1.93 0.70 1.97 3.87 8.58 To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. Table 5: Change in salivary cortisol between baseline and 60 minutes in healthy volunteers, adrenal insufficient patients, and cirrhotic patients in the 250 𝜇g ACTH test. Delta salivary cortisol (𝜇g/dL) Mean Median Std Min Q1 Q3 Max Group Healthy 1.79 1.67 0.59 0.99 1.38 2.03 3.61 Adrenal insufficiency -0.00 0.00 0.18 -0.72 -0.02 0.06 0.35 Cirrhosis 1.53 1.33 0.99 0.08 0.80 1.98 4.65 To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. Healthy Adrenal insufficiency 50 40 30 cortisol 20 10 0 0 30 60 90 0 30 60 90 time serum cortisol salivary cortisol Figure 1: Serum and salivary cortisol (𝜇g/dL) in the 250 𝜇g ACTH test. To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. International Journal of Endocrinology 5 Table 6: Pearson correlation (r) between serum and salivary cortisol concentrations in healthy volunteers and adrenal insufficient patients at different timepoints during the ACTH test. HV SaC AI SaC T0 T30 T60 T90 T0 T30 T60 T90 T0 0.83a - - - - - - - T30 - 0.67a - - - - - - HV TSC T60 - - 0.54b - - - - - T90 - - - 0.62a - - - - T0 - - - - 0.14c - - - T30 - - - - - 0.49b - - AI TSC T60 - - - - - - 0.71a - T90 - - - - - - - 0.79a HV, healthy volunteers; AI, adrenal insufficiency; TSC, total serum cortisol; SaC, salivary cortisol; T0 , baseline; T30 , 30 min after administration; T60 , 60 min after administration; T90 , 90 min after administration. a p <0.0001, b p <0.01, and c p >0.05. Table 7: Number of cirrhotic patients meeting at least one criterion for the diagnosis of AI based on serum cortisol and salivary cortisol thresholds during 250 𝜇g ACTH test (N=39). Criterion AI based on TSC AI based on SaC AI defined from T0 , n 13a 0d AI defined from T60 , n 9b 9e AI defined from the difference between 0 and 60 min, n 10c 12f Total number of patients reaching at least one of the three criteria of adrenal 19 (48.7%) 12 (30.8%) insufficiency, n (%) AI, adrenal insufficiency; TSC, total serum cortisol; SaC, salivary cortisol; T0 , baseline; T60 , 60 min after administration. To convert cortisol micrograms per deciliter to nanomoles per liter, multiply by 27.6. a T0 TSC <9 𝜇g/dL, b T60 TSC <18 𝜇g/dL, c ΔTSC<9 𝜇g/dL. d T0 SaC<0.08 𝜇g/dL, e T60 SaC<1.43 𝜇g/dL, f ΔSaC<1 𝜇g/dL. Table 6 reports the correlations between SaC and TSC Based on these results, we classified cirrhotic patients into concentrations at different timepoints during the ACTH test. three groups: No AI (n=19), AI based on TSC and SaC criteria In AI patients, SaC and TSC correlated except at baseline, (n=12), and AI based on TSC but not SaC criteria (n=8). and the strength of the correlations increased over time. By Comparing the characteristics of the patients in these groups, contrast, in healthy patients, SaC and TSC correlated strongly we found that the No-AI group had a higher proportion of at all timepoints. women (X2 =17.14; p<0.001), more patients with non-alcohol- related cirrhosis (although only 2 patients had non-alcohol- 3.2. Cirrhotic Subjects. Mean values of TSC and SaC at the related cirrhosis), and a greater-than-expected proportion different timepoints in the ACTH test in cirrhotic patients of Child A and Child B patients (X2 =20.10; p=0.0005); are reported in Tables 2 and 4, respectively; mean values of both AI groups had a greater-than-expected proportion of ΔTSC and ΔSaC are reported in Tables 3 and 5, respectively. patients with alcohol-related cirrhosis (X2 =25.29; p=0.0014). Based on the results for healthy subjects and AI patients, We found no significant differences between the three groups we selected the following cutoffs for the diagnosis of AI in in age at diagnosis of cirrhosis, current MELD, blood cirrhotic patients SaC-T0 < 0.08 𝜇g/dL (2.2 nmol/L) or SaC- albumin, prealbumin, HDL or LDL cholesterol, creatinine, T60 < 1.43 𝜇g/dl (39.5 nmol/L) or ΔSaC<1 𝜇g/dl (27.6 nmol/L). triglycerides, AST, ALT, bilirubin, prothrombin time, INR, or Table 7 reports the numbers of cirrhotic patients that reasons for admission.Cirrhotic subjects with AI diagnosed met each criterion for the diagnosis of AI with each method. according to SaC levels were treated with glucocorticoid Comparing the results of using the SaC thresholds versus replacement therapy. the established TSC thresholds to diagnose AI in cirrhotic patients, we found 19 patients met at least one TSC criterion of 4. Discussion AI and 12 patients met at least one SaC criterion; however, this difference in frequency did not reach statistical significance. To determine whether SaC can be used for diagnosing AI in The criteria for AI according to both of the two methods noncritical cirrhotic patients, we established reference values were met by 11 (28.2%) patients; 8 (20.5%) met only the TSC for SaC at T0 and T60 and for ΔSaC (T0 -T60 ) based on SaC and criteria, and 1 (2.6%) met only the SaC criteria. TSC findings in normal subjects and patients with known AI. 6 International Journal of Endocrinology We found that SaC can be very useful for diagnosis of AI in The 30.8% SaC-determined prevalence in our noncritical cirrhotic patients. cirrhotic patients is higher than the 9.1% found by Galbois Determining SaC is a quick, easy, noninvasive technique et al. [32]; it is also higher than the 19% prevalence calculated used since the early 1980s, when SaC was discovered to by Fede et al. [33] by measuring FC after a 1 𝜇g ACTH test. be an excellent indicator of plasma FC concentration [21, As in other studies, we found that TSC overestimated the 22]. Various authors have since studied TSC and SaC in prevalence of AI (48.7% in our study, 33% in Galbois et al. different circumstances in which cortisol-binding globulin [32], and 34% in Fede et al. [33]). The higher prevalence of is altered (oral contraception, pregnancy, and cirrhosis) [10, AI in our study might be due to greater liver disease severity 23]. In 2009, Deutschbein et al. [23] compared basal SaC (85% Child B or C); we found that severity was associated and basal TSC to the insulin tolerance test in 77 patients with a higher risk of AI. We also found that sex and cirrhosis with hypothalamic-pituitary disease, concluding that both etiology were associated with AI frequency. approaches enabled a highly specific diagnosis, obviating Several studies identified ascites, low HDL-cholesterol, insulin tolerance testing in about one-fourth of cases. In and liver disease severity as risk factors for AI [32, 34–36]; 2012, they found basal SaC<0.11 𝜇g/dL (3.0 nmol/L) had other reported risk factors include low cortisol-binding glob- 97% specificity and 40% sensitivity for AI, enabling correct ulin [33], higher MELD score [37], and lower serum albumin classification in 26% [24]. By contrast, Ceccato et al. [25] [32]. Galbois et al. [32] reported that hypoalbuminemia was concluded that unstimulated SaC<0.09 𝜇g/dL (2.5 nmol/L) the main reason for discrepancies between TSC and SaC distinguished AI patients from healthy subjects with 97.1% assessments of AI, suggesting that a lower threshold of 25 sensitivity and 93.3% specificity. Recently, Langelaan et al. g/L could be used to identify patients who could benefit [26] recommended a new diagnostic algorithm to diagnose from a SaC assessment. When we analyzed the subgroup of AI, with early morning SaC>0.21 𝜇g/dL (5.8 nmol/L) ruling patients with albumin<25 g/L, we found no differences in the out AI and <0.04 𝜇g/dL (1.1 nmol/L) diagnosing AI; these cut- prevalence of AI measured by SaC or TSC, probably because offs enabled 34% of patients to be diagnosed without ACTH few (8/39) patients had albumin<25 g/L. stimulation. In our study, SaC-T0 < 0.08 𝜇g/dL (2.2 nmol/L) Our study has several limitations. We calculated disease- (the lowest SaC-T0 value in normal subjects) diagnosed specific thresholds of SaC measured by electrochemilumi- 7.7% of patients with AI; the differences in cutoffs might be nescence immunoassay, the method routinely used in our due to differences in laboratory methods for determining hospital. Liquid chromatography tandem-mass spectrometry SaC. is more specific but requires expensive equipment [38]. In 1988, Laudat et al. [27] found no overlap in SaC-T60 Despite the risk of overestimating SaC, potential cross- after 250 𝜇g ACTH between 58 healthy volunteers and 21 reactivity with other steroids, and different results obtained subjects with AI; discrepancies between SaC and TSC in 8 with different types of analyzers [39, 40], we consider the use patients with AI were attributed to thyroid hormones and of electrochemiluminescence immunoassay justified because psychotropic agents. Various studies have since examined it is the method most commonly used in clinical practice. correlations between TSC and SaC during 1 𝜇g or 250 𝜇g Moreover, we did not measure FC, because doing so is ACTH tests [16, 28–30]. Correlations between TSC and SaC complex, expensive, and uncommon in clinical practice; are good; the issue is choosing the optimal SaC cutoff for likewise, we did not measure cortisol-binding globulin, so we AI. Methodological differences among studies make com- could not calculate FC with Coolens’ equation. parisons difficult. Some studies included healthy volunteers In diagnostic accuracy studies, how eligible subjects are and patients with known AI, whereas others included sub- identified and recruited is important. We included only jects with suspected AI. Moreover, some used 1 𝜇g ACTH, hemodynamically stable cirrhotic patients, so our findings whereas others used 250 𝜇g. Finally, different studies used cannot be extrapolated to cirrhotic patients with sepsis or different methods to measure SaC and TSC, and the normal septic shock. On the other hand, we determined cutoffs response to ACTH tests is assay-specific [31]. Thus, despite from our findings in patients with known AI and healthy the available data, clinicians face uncertainty in choosing volunteers with low probability of AI, adding strength to cutoffs. our results. Other studies derived cutoffs from findings in We compared SaC and TSC before and after adminis- patients with suspected AI, and that design could influence tering 250 𝜇g ACTH in healthy subjects and patients with the spectrum of disease in included patients [41]. Moreover, known AI to calculate reference values for basal SaC, peak test sensitivity is usually higher in studies with patients with SaC at T60 , and ΔSaC (T0 -T60 ). We used these references to more advanced stages of the target condition [42]. determine the prevalence of AI in a cohort of decompensated Measuring SaC has some methodological limitations. cirrhotic patients and compared the results with those found SaC’s concentration is 30-fold lower than TSC’s. Additionally, using established TSC cutoffs. Whereas the established TSC SaC levels are affected by salivary 11𝛽-hydroxysteroid dehy- cutoffs classified 48.7% of the cirrhotic patients as AI, our drogenase type 2. In 2010, Perogamvros et al. [43] suggested SaC cutoffs classified only 30.8% as AI. The difference was using salivary cortisone better reflects FC after adrenal not statistically significant, probably due to the low number stimulation. Cornes et al. [40] report that concentrations of of cirrhotic patients, but it shows a clear tendency. At least salivary cortisone are three times higher than those of SaC one SaC criterion and one TSC criterion for AI were met in 11 and have a closer linear correlation with serum FC. Debono (28.2%); 8 (20.5%) had AI according to TSC but not according et al. [44] also suggested that salivary cortisone may be the to SaC, and 1 patient had AI according to SaC but not TSC. preferred analyte for noninvasive measurement of FC. International Journal of Endocrinology 7 5. Conclusions [3] B. Zietz, G. Lock, B. Plach et al., “Dysfunction of the hypothalamic-pituitary-glandular axes and relation to Child- We establish method-specific reference cutoffs of SaC and Pugh classification in male patients with alcoholic and virus- ΔSaC to determine AI during the 250 𝜇g ACTH stimulation related cirrhosis,” European Journal of Gastroenterology & Hep- test. SaC is more accurate than TSC for assessing adrenal atology, vol. 15, no. 5, pp. 495–501, 2003. function with the 250 𝜇g ACTH stimulation test in non- [4] S. R. Bornstein, B. Allolio, W. Arlt et al., “Diagnosis and critical cirrhotic patients. The prevalence AI in our cirrhotic treatment of primary adrenal insufficiency: An endocrine group was higher than in other studies. However, it should society clinical practice guideline,” The Journal of Clinical be taken into account that we did not measure CBG levels. Endocrinology & Metabolism, vol. 101, no. 2, pp. 364–389, 2016. Further studies with more cases are necessary to establish [5] R. Kazlauskaite, A. T. Evans, C. V. Villabona et al., “Corti- SaC reference values to correctly classify patients with AI and cotropin tests for hypothalamic-pituitary-adrenal insufficiency: avoid unnecessary cortisol replacement treatment. A metaanalysis,” The Journal of Clinical Endocrinology & Metabolism, vol. 93, no. 11, pp. 4245–4253, 2008. [6] W. Rosner, “The functions of corticosteroid-binding glob- Data Availability ulin and sex hormone-binding globulin: Recent advances,” Endocrine Reviews, vol. 11, no. 1, pp. 80–91, 1990. The data used to support the findings of this study are [7] J. D. Baxter and P. H. Forsham, “Tissue effects of glucocorti- available from the corresponding author upon request. coids,” American Journal of Medicine, vol. 53, no. 5, pp. 573–589, 1972. [8] J. F. Dunn, B. C. Nisula, and D. Rodbard, “Transport of Ethical Approval steroid hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding glob- Our institution’s ethics committee approved this prospective ulin in human plasma,” The Journal of Clinical Endocrinology & study. Metabolism, vol. 53, no. 1, pp. 58–68, 1981. [9] R. P. Vincent, F. E. Etogo-Asse, T. Dew, W. Bernal, J. Consent Alaghband-Zadeh, and C. W. Le Roux, “Serum total cortisol and free cortisol index give different information regarding the All subjects provided written informed consent. hypothalamus-pituitary-adrenal axis reserve in patients with liver impairment,” Annals of Clinical Biochemistry, vol. 46, no. 6, pp. 505–507, 2009. Disclosure [10] D. Suri, J. Moran, J. U. Hibbard, K. Kasza, and R. E. Weiss, “Assessment of adrenal reserve in pregnancy: defining the An earlier version of the study has been presented as a poster normal response to the adrenocorticotropin stimulation test,” in Cirrhosis: Clinical aspects. https://www.sciencedirect.com/ The Journal of Clinical Endocrinology & Metabolism, vol. 91, no. science/article/abs/pii/S0168827816003056. 10, pp. 3866–3872, 2006. [11] H. Raff, S. Brock, and J. W. Findling, “Cosyntropin-stimulated salivary cortisol in hospitalized patients with hypoproteinemia,” Conflicts of Interest Endocrine Journal, vol. 34, no. 1-3, pp. 68–74, 2008. [12] M. Rauschecker, S. B. Abraham, B. S. Abel et al., “Cosyntropin- The authors have no conflicts of interest regarding this study. stimulated serum free cortisol in healthy, adrenally insuffi- cient, and mildly cirrhotic populations,” The Journal of Clinical Endocrinology & Metabolism, vol. 101, no. 3, pp. 1075–1081, 2016. Authors’ Contributions [13] N. Molenaar, A. B. J. Groeneveld, and M. F. C. de Jong, “Three Lara Albert and Joaquı́m Profitós contributed equally to this calculations of free cortisol versus measured values in the critically ill,” Clinical Biochemistry, vol. 48, no. 16-17, pp. 1053– work. 1058, 2015. [14] B. M. Arafah, F. J. Nishiyama, H. Tlaygeh, and R. Hejal, Acknowledgments “Measurement of salivary cortisol concentration in the assess- ment of adrenal function in critically ill subjects: A surrogate This study was financed by a grant from Fundació Parc Taulı́: marker of the circulating free cortisol,” The Journal of Clinical CIR [2012/22]. Endocrinology & Metabolism, vol. 92, no. 8, pp. 2965–2971, 2007. [15] H. Raff, “Utility of salivary cortisol measurements in Cushing’s syndrome and adrenal insufficiency,” The Journal of Clinical References Endocrinology & Metabolism, vol. 94, no. 10, pp. 3647–3655, 2009. [1] J. O’Beirne, M. Holmes, B. Agarwal et al., “Adrenal insufficiency [16] J. G. Lewis, “Steroid analysis in saliva: an overview,” The Clinical in liver disease - What is the evidence?” Journal of Hepatology, Biochemist Reviews, vol. 27, no. 3, pp. 139–146, 2006. vol. 47, no. 3, pp. 418–423, 2007. [17] P. E. Marik, S. M. Pastores, D. Annane et al., “Recommendations [2] M.-H. Tsai, Y.-S. Peng, and Y.-C. Chen, “Adrenal insufficiency for the diagnosis and management of corticosteroid insuffi- in patients with cirrhosis, severe sepsis and septic shock,” ciency in critically ill adult patients: consensus statements from Hepatology, vol. 43, no. 4, pp. 673–681, 2006. an international task force by the American College of Critical 8 International Journal of Endocrinology Care Medicine,” Critical Care Medicine, vol. 36, no. 6, pp. 1937– [31] M. C. Klose, M. Lange, A. K. Rasmussen et al., “Factors influenc- 1949, 2008. ing the adrenocorticotropin test: Role of contemporary cortisol [18] R. I. Dorin, C. R. Qualls, and L. M. Crapo, “Diagnosis of adrenal assays, body composition, and oral contraceptive agents,” The insufficiency,” Annals of Internal Medicine, vol. 139, no. 3, pp. Journal of Clinical Endocrinology & Metabolism, vol. 92, no. 4, 194–204, 2003. pp. 1326–1333, 2007. [19] I. Lopez Schmidt, H. Lahner, K. Mann, and S. Petersenn, “Diag- [32] A. Galbois, M. Rudler, J. Massard et al., “Assessment of adrenal nosis of adrenal insufficiency: Evaluation of the corticotropin- function in cirrhotic patients: Salivary cortisol should be pre- releasing hormone test and basal serum cortisol in comparison ferred,” Journal of Hepatology, vol. 52, no. 6, pp. 839–845, 2010. to the insulin tolerance test in patients with hypothalamic- [33] G. Fede, L. Spadaro, T. Tomaselli et al., “Comparison of total pituitary-adrenal disease,” The Journal of Clinical Endocrinology cortisol, free cortisol, and surrogate markers of free cortisol & Metabolism, vol. 88, no. 9, pp. 4193–4198, 2003. in diagnosis of adrenal insufficiency in patients with stable [20] R. F. Vining, R. A. McGinley, J. J. Maksvytis, and K. Y. Ho, cirrhosis,” Clinical Gastroenterology and Hepatology, vol. 12, no. “Salivary cortisol: a better measure of adrenal cortical function 3, pp. 504–e8, 2014. than serum cortisol,” Annals of Clinical Biochemistry, vol. 20, [34] P. E. Marik, T. Gayowski, and T. E. Starzl, “The hepatoadrenal part 6, pp. 329–335, 1983. syndrome: a common yet unrecognized clinical condition,” [21] U. Teruhisa, H. Ryoji, I. Taisuke, S. Tatsuya, M. Fumihiro, and Critical Care Medicine, vol. 33, no. 6, pp. 1254–1259, 2005. S. Tatsuo, “Use of saliva for monitoring unbound free cortisol [35] P. E. Marik, “Adrenal-exhaustion syndrome in patients with levels in serum,” Clinica Chimica Acta, vol. 110, no. 2-3, pp. 245– liver disease,” Intensive Care Medicine, vol. 32, no. 2, pp. 275– 253, 1981. 280, 2006. [22] B. J. Nolan, J. Sorbello, N. Brown, G. Dimeski, and W. J. Inder, [36] P. H. J. Van Der Voort, R. T. Gerritsen, A. J. Bakker, E. C. “Characterization of the serum and salivary cortisol response to Boerma, M. A. Kuiper, and L. De Heide, “HDL-cholesterol level the intravenous 250 𝜇g ACTH1-24 stimulation test,” Endocrine and cortisol response to synacthen in critically ill patients,” Journal, vol. 59, no. 3, pp. 520–528, 2018. Intensive Care Medicine, vol. 29, no. 12, pp. 2199–2203, 2003. [23] T. Deutschbein, N. Unger, K. Mann, and S. Petersenn, “Diag- [37] T. Thevenot, S. Borot, A. Remy-Martin et al., “Assessment of nosis of secondary adrenal insufficiency: Unstimulated early adrenal function in cirrhotic patients using concentration of morning cortisol in saliva and serum in comparison with the serum-free and salivary cortisol,” Liver International, vol. 31, no. insulin tolerance test,” Hormone and Metabolic Research, vol. 41, 3, pp. 425–433, 2011. no. 11, pp. 834–839, 2009. [38] A. E. Taylor, B. Keevil, and I. T. Huhtaniemi, “Mass spectrome- [24] T. Deutschbein, M. Broecker-Preuss, J. Flitsch et al., “Salivary try and immunoassay: How to measure steroid hormones today cortisol as a diagnostic tool for Cushing’s syndrome and adrenal and tomorrow,” European Journal of Endocrinology, vol. 173, no. insufficiency: Improved screening by an automatic immunoas- 2, pp. D1–D12, 2015. say,” European Journal of Endocrinology, vol. 166, no. 4, pp. 613– [39] M. Dušková, K. Šimůnková, J. Vı́tků et al., “A comparison 618, 2012. of salivary steroid levels during diagnostic tests for adrenal [25] F. Ceccato, M. Barbot, M. Zilio et al., “Performance of sali- insufficiency,” Prague Medical Report, vol. 117, no. 1, pp. 18–33, vary cortisol in the diagnosis of Cushing’s syndrome, adrenal 2016. incidentaloma, and adrenal insufficiency,” European Journal of [40] M. P. Cornes, H. L. Ashby, Y. Khalid, H. N. Buch, C. Ford, Endocrinology, vol. 169, no. 1, pp. 31–36, 2013. and R. Gama, “Salivary cortisol and cortisone responses to [26] M. L. P. Langelaan, J. M. H. Kisters, M. M. Oosterwerff, tetracosactrin (synacthen),” Annals of Clinical Biochemistry, vol. and A.-K. Boer, “Salivary cortisol in the diagnosis of adrenal 52, no. 5, pp. 606–610, 2015. insufficiency: Cost efficient and patient friendly,” Endocrine [41] P. M. Bossuyt, J. B. Reitsma, D. E. Bruns et al., “STARD 2015: An Connections, vol. 7, no. 4, pp. 560–566, 2018. updated list of essential items for reporting diagnostic accuracy studies,” BMJ, vol. 351, p. h5527, 2015. [27] M. H. Laudat, S. Cerdas, C. Fournier, D. Guiban, B. Guilhaume, and J. P. Luton, “Salivary cortisol measurement: A practical [42] D. F. Ransohoff and A. R. Feinstein, “Problems of spectrum approach to assess pituitary-adrenal function,” The Journal of and bias in evaluating the efficacy of diagnostic tests,” The New Clinical Endocrinology & Metabolism, vol. 66, no. 2, pp. 343– England Journal of Medicine, vol. 299, no. 17, pp. 926–930, 1978. 348, 1988. [43] I. Perogamvros, B. G. Keevil, D. W. Ray, and P. J. Trainer, [28] Y. Marcus-Perlman, K. Tordjman, Y. Greenman et al., “Low- “Salivary cortisone is a potential biomarker for serum free dose ACTH (1 𝜇g) salivary test: A potential alternative to the cortisol,” The Journal of Clinical Endocrinology & Metabolism, classical blood test,” Clinical Endocrinology, vol. 64, no. 2, pp. vol. 95, no. 11, pp. 4951–4958, 2010. 215–218, 2006. [44] M. Debono, R. F. Harrison, M. J. Whitaker et al., “Salivary cor- [29] T. Deutschbein, N. Unger, K. Mann, and S. Petersenn, “Diag- tisone reflects cortisol exposure under physiological conditions nosis of secondary adrenal insufficiency in patients with and after hydrocortisone,” The Journal of Clinical Endocrinology hypothalamic-pituitary disease: Comparison between serum & Metabolism, vol. 101, no. 4, pp. 1469–1477, 2016. and salivary cortisol during the high-dose short synacthen test,” European Journal of Endocrinology, vol. 160, no. 1, pp. 9–16, 2009. [30] I. Perogamvros, L. J. Owen, B. G. Keevil, G. Brabant, and P. J. Trainer, “Measurement of salivary cortisol with liquid chromatography-tandem mass spectrometry in patients under- going dynamic endocrine testing,” Clinical Endocrinology, vol. 72, no. 1, pp. 17–21, 2010.