Advertisement

Including Relative Adrenal Insufficiency in Definition and Classification of Acute-on-Chronic Liver Failure

Published:October 04, 2019DOI:https://doi.org/10.1016/j.cgh.2019.09.035

      Background & Aims

      Relative adrenal insufficiency (RAI) is defined by insufficient production of cortisol relative to organ demand. RAI is observed frequently in hospitalized patients with cirrhosis, but there is disagreement over the clinical effects of RAI in these patients. We evaluated the prevalence and the clinical effects of RAI in hospitalized patients with cirrhosis.

      Methods

      We performed a prospective study of 160 patients admitted to a hospital in Italy for acute decompensation of cirrhosis from May 2011 through September 2016. Patients were followed up until death, liver transplantation, or a maximum of 90 days. Serum and salivary levels of cortisol were measured before and after a 1-hour Short Synacthen Test. A diagnosis of RAI was given to patients with an increase in serum cortisol of less than 9 μg/dL, after Synacthen administration, in patients with baseline serum levels of cortisol less than 35 μg/dL. We collected blood samples before the Synacthen test and analyzed them for blood cell counts, liver and renal function, levels of C-reactive protein, and lipid profiles (total cholesterol, high-density lipoprotein cholesterol, apolipoprotein-A1).

      Results

      A diagnosis of RAI was made for 78 patients (49%). Age (odds ratio [OR], 0.95; P = .030), number of leukocytes (OR, 3.10; P = .006), and levels of high-density lipoprotein cholesterol (OR, 0.30; P = .039) were associated independently with RAI. Patients with RAI had a significantly higher risk of developing bacterial infections (hazard ratio [HR], 1.60; P = .038), sepsis (HR, 2.95; P = .001), septic shock (HR, 4.94; P = .038), new organ failures (HR, 2.45; P = .014), and acute-on-chronic liver failure (HR, 2.27; P = .037) than patients without RAI. RAI was associated independently with death within 90 days of diagnosis (subdistribution HR, 4.83; P = .001). Patients with RAI and mild renal dysfunction or hepatic encephalopathy had no significant difference in cumulative incidence of 28-day mortality vs patients with acute-on-chronic liver failure grade 1 (25% vs 22%).

      Conclusions

      We found RAI to occur in almost half of patients admitted to a hospital for acute decompensation of cirrhosis. RAI was associated with a deficit of substrates for steroidogenesis and an increase in markers of inflammation. Patients with RAI have a high risk of developing sepsis, septic shock, organ failure, and death within 90 days. RAI has similar prognostic value to nonrenal organ failures.

      Keywords

      Abbreviations used in this paper:

      ACLF (acute-on-chronic liver failure), ACTH (adrenocorticotropic hormone), AKI (acute kidney injury), CRP (C-reactive protein), EASL-CLIF (European Association for the Study of the Liver-Chronic Liver Failure), HDL (high-density lipoprotein), OR (odds ratio), RAI (relative adrenal insufficiency), sHR (subdistribution hazard ratio), SST (Short Synacthen Test)
      See editorial on page 1040.

       Background

      Relative adrenal insufficiency (RAI, insufficient production of cortisol relative to organ demand) occurs frequently in hospitalized patients with cirrhosis. We evaluated the prevalence and the clinical effects of RAI in 160 patients hospitalized for acute decompensation of cirrhosis.

       Findings

      RAI occurs in almost half of patients admitted to a hospital for acute decompensation of cirrhosis, and is associated with a deficit of substrates for steroidogenesis and an increase in markers of inflammation. Patients with RAI are at high risk for sepsis, septic shock, organ failure, and death within 90 days.

       Implications for patients care

      RAI can be identified easily with the Short Synacthen Test and has similar prognostic value to nonrenal organ failure.
      The hypothalamic-pituitary-adrenal axis plays a key role in modulating the body’s response to stress. During acute illness, the activation of the hypothalamic-pituitary-adrenal axis leads to the production of cortisol in the adrenal grands, which contributes to maintain cellular and organ homeostasis.
      • Cooper M.S.
      • Stewart P.M.
      Corticosteroid insufficiency in acutely ill patients.
      Relative adrenal insufficiency (RAI) is a syndrome characterized by inadequate glucocorticoid activity relative to illness severity.
      • Cooper M.S.
      • Stewart P.M.
      Corticosteroid insufficiency in acutely ill patients.
      ,
      • Marik P.E.
      • Pastores S.M.
      • Annane D.
      • et al.
      Recommendations for the diagnosis and management of corticosteroid insufficiency in critically ill adult patients: consensus statements from an international task force by the American College of Critical Care Medicine.
      In critically ill patients, RAI is better known as critical illness–related corticosteroid insufficiency and it has been associated with high short-term mortality.
      • Annane D.
      • Sébille V.
      • Troché G.
      • et al.
      A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin.
      In critically ill patients with cirrhosis (mainly with severe sepsis/septic shock), RAI was observed in 40% to 77% of patients and was found to be associated with poor in-hospital survival, refractory shock, and renal failure.
      • Fernández J.
      • Escorsell A.
      • Zabalza M.
      • et al.
      Adrenal insufficiency in patients with cirrhosis and septic shock: effect of treatment with hydrocortisone on survival.
      • Tsai M.-H.
      • Peng Y.-S.
      • Chen Y.-C.
      • et al.
      Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock.
      • Arabi Y.M.
      • Aljumah A.
      • Dabbagh O.
      • et al.
      Low-dose hydrocortisone in patients with cirrhosis and septic shock: a randomized controlled trial.
      • Tsai M.-H.
      • Huang H.-C.
      • Peng Y.-S.
      • et al.
      Critical illness–related corticosteroid insufficiency in cirrhotic patients with acute gastroesophageal variceal bleeding: risk factors and association with outcome*.
      Data collected in non–critically ill patients provided conflicting results, both on the prevalence (range, 7%–60%) and clinical impact of RAI in these patients.
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      • Tan T.
      • Chang L.
      • Woodward A.
      • et al.
      Characterising adrenal function using directly measured plasma free cortisol in stable severe liver disease.
      • Fede G.
      • Spadaro L.
      • Tomaselli T.
      • et al.
      Assessment of adrenocortical reserve in stable patients with cirrhosis.
      • Thevenot T.
      • Borot S.
      • Remy-Martin A.
      • et al.
      Assessment of adrenal function in cirrhotic patients using concentration of serum-free and salivary cortisol.
      • Acevedo J.
      • Fernández J.
      • Prado V.
      • et al.
      Relative adrenal insufficiency in decompensated cirrhosis: relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death.
      • Graupera I.
      • Pavel O.
      • Hernandez-Gea V.
      • et al.
      Relative adrenal insufficiency in severe acute variceal and non-variceal bleeding: influence on outcomes.
      • Risso A.
      • Alessandria C.
      • Mezzabotta L.
      • et al.
      Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: relationship and effect on survival.
      • Fede G.
      • Spadaro L.
      • Tomaselli T.
      • et al.
      Comparison of total cortisol, free cortisol, and surrogate markers of free cortisol in diagnosis of adrenal insufficiency in patients with stable cirrhosis.
      • Jang J.Y.
      • Kim T.Y.
      • Sohn J.H.
      • et al.
      Relative adrenal insufficiency in chronic liver disease: its prevalence and effects on long-term mortality.
      This discrepancy resulted from different methods used to define RAI and low sample size; further studies are needed to clarify the prevalence and clinical impact of RAI in patients with cirrhosis. Thus, RAI is not yet a clear indication for treatment with steroids in these patients.
      The pathogenesis of RAI in patients with cirrhosis is not clear, but low levels of substrates for steroidogenesis (cholesterol), systemic inflammation, and adrenal hypoperfusion owing to circulatory dysfunction potentially could be implicated in the development of RAI.
      • Fede G.
      • Spadaro L.
      • Tomaselli T.
      • et al.
      Adrenocortical dysfunction in liver disease: a systematic review.
      Interestingly, in patients with an acute decompensation of cirrhosis, systemic inflammation and circulatory dysfunction are also the main drivers of organ failures,
      • Clària J.
      • Stauber R.E.
      • Coenraad M.J.
      • et al.
      Systemic inflammation in decompensated cirrhosis: characterization and role in acute-on-chronic liver failure.
      ,
      • Bernardi M.
      • Moreau R.
      • Angeli P.
      • et al.
      Mechanisms of decompensation and organ failure in cirrhosis: from peripheral arterial vasodilation to systemic inflammation hypothesis.
      which are the hallmark of acute-on-chronic liver failure (ACLF), a syndrome associated with poor short-term survival in these patients.
      • Moreau R.
      • Jalan R.
      • Gines P.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      However, RAI has not been studied in the perspective of ACLF.
      The aims of this study were as follows: (1) to assess the prevalence of RAI in patients hospitalized for an acute decompensation of cirrhosis; (2) to explore factors involved in the pathogenesis of RAI; (3) to evaluate the impact of RAI on clinical outcomes in these patients; and (4) to evaluate the potential impact of RAI on the definition and grading of ACLF.

      Patients and Methods

       Patients

      Patients admitted to the University Hospital of Padua in Italy from May 2011 to September 2016 were screened prospectively for participating in the study. Inclusion criteria were as follows: (1) diagnosis of cirrhosis according to clinical, biochemical, ultrasonographic, endoscopic, and/or histologic findings; (2) at least 18 years of age; and (3) acute decompensation of cirrhosis (ascites, bacterial infections, gastrointestinal bleeding, hepatic encephalopathy, and so forth). Exclusion criteria were as follows: (1) admission for planned procedures; (2) time elapsed from the hospital admission of more than 24 hours; (3) hepatocellular carcinoma beyond the Milan criteria; (4) severe extrahepatic diseases (congestive heart failure stage New York Heart Association class ≥2, chronic obstructive pulmonary disease stage Global initiative for chronic Obstructive Lung Disease ≥2; chronic kidney disease requiring renal replacement therapy); (5) septic shock; (6) previous transplant; (7) patients with known adrenal, hypothalamic, and/or pituitary gland diseases; (8) treatment with steroids and/or other immunosuppressive drugs; (9) parenteral feeding; and (10) extrahepatic malignancies. All patients signed informed consent, and the study followed the Declaration of Helsinki and was approved by the local Ethics committee.

       Study Design

      Medical history, demographic data, and clinical data were collected at inclusion in the study. Adrenal function was investigated performing the Short Synacthen Test (SST). Synthetic adrenocorticotropic hormone (ACTH, 250 μg, Synacthen; Novartis Pharma, Basel, Switzerland) was given intravenously between 8:00 and 9:00 AM after at least 8 hours of fasting and bed rest. Serum and saliva samples were collected before and 60 minutes after the administration of Synacthen to measure cortisol levels. The patients brushed their teeth at least 30 minutes before saliva collection. Salivary samples were collected to check for the absence of blood in the mouth. Saliva was collected using Salivette devices (Sarstedt, Nümbrecht, Germany), as previously described in detail.
      • Antonelli G.
      • Ceccato F.
      • Artusi C.
      • et al.
      Salivary cortisol and cortisone by LC–MS/MS: validation, reference intervals and diagnostic accuracy in Cushing’s syndrome.
      In addition, plasma samples for the assessment of inflammatory cytokines were collected before the administration of Synacthen. Plasma and saliva samples were stored at -80°C. Tumor necrosis factor-α, interleukin 1β, interleukin 6, and interleukin 10 concentrations were measured by an enzyme-linked immunosorbent assay kit (RayBiotech, Norcross, GA) with a fluorometric assay.
      Serum cortisol levels were measured by a commercial competitive chemiluminescence enzyme immunoassay (Immulite 2000; Siemens, Munich, Germany).
      Salivary cortisol and cortisone levels were measured by a homemade liquid chromatography–tandem mass spectrometry, with automated sample preparation, as previously described and reported in detail in the Supplementary Methods section.
      • Antonelli G.
      • Ceccato F.
      • Artusi C.
      • et al.
      Salivary cortisol and cortisone by LC–MS/MS: validation, reference intervals and diagnostic accuracy in Cushing’s syndrome.
      Laboratory tests including whole blood count, liver and renal function, C-reactive protein (CRP), and lipid profiles (total cholesterol, high-density lipoprotein [HDL] cholesterol, apolipoprotein-A1) were performed immediately before the SST test.
      Patients with RAI were not treated with steroids because there is no clear evidence that steroids are beneficial in these patients.
      • Angeli P.
      • Bernardi M.
      • Villanueva C.
      • et al.
      EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis.
      Patients were prospectively followed up until death, liver transplantation, or for a maximum of 90 days. Data on the development of complications of cirrhosis (hepatic encephalopathy, gastrointestinal bleeding), bacterial infections, sepsis and/or septic shock, acute kidney injury (AKI), organ failures, and ACLF were collected prospectively.

       Definitions

      RAI was defined by an increase in serum total cortisol after SST less than 9 μg/dL in patients with basal serum total cortisol less than 35 μg/dL. This is the definition recommended by a task force of experts in critically ill patients.
      • Marik P.E.
      • Pastores S.M.
      • Annane D.
      • et al.
      Recommendations for the diagnosis and management of corticosteroid insufficiency in critically ill adult patients: consensus statements from an international task force by the American College of Critical Care Medicine.
      In addition, it is not affected by the low levels of albumin and transcortin observed in cirrhotic patients that may cause a false low concentration of total blood cortisol leading to an overestimation of RAI. Organ failures and ACLF were defined according to the European Association for the Study of the Liver-Chronic Liver Failure (EASL-CLIF) consortium criteria
      • Moreau R.
      • Jalan R.
      • Gines P.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      according to an amendment to the study protocol (details are reported in the Supplementary Methods section). AKI was defined according to the International Club of Ascites AKI criteria.
      • Angeli P.
      • Gines P.
      • Wong F.
      • et al.
      Diagnosis and management of acute kidney injury in patients with cirrhosis: revised consensus recommendations of the International Club of Ascites.
      Hyponatremia was defined as a serum sodium level less than 130 mmol/L. Sepsis and septic shock were defined according to Sepsis-3 criteria, as suggested for patients with cirrhosis.
      • Piano S.
      • Bartoletti M.
      • Tonon M.
      • et al.
      Assessment of Sepsis-3 criteria and quick SOFA in patients with cirrhosis and bacterial infections.
      Since the Sepsis-3 criteria were published in 2016, the assessment of sepsis occurrence during the follow-up period was performed retrospectively using the collected data (Supplementary Methods).

       Statistical Analysis

      Continuous variables were reported as means ± SD or as medians with interquartile range according to normal or non-normal distribution, respectively. They were compared using the Student t test or the Mann–Whitney U test, respectively. Categoric variables were reported as proportions and compared with the chi-square test or the Fisher exact test. Variables associated with RAI with a P value less than .1 in the univariate analysis were included in a multivariate stepwise logistic regression analysis with backward elimination (entry P < .05; drop P > .1). The odds ratios (OR) and their 95% CI were calculated. The cumulative incidence of 90-day mortality was calculated considering liver transplantation as a competing event and compared using the Gray test. Univariate and multivariate analyses of mortality were performed using the proportional hazards model for the subdistribution of competing risks proposed by Fine and Gray,
      • Fine J.P.
      • Gray R.J.
      A proportional hazards model for the subdistribution of a competing risk.
      and results were expressed as a P value, subdistribution hazard ratios (sHR), and their 95% CIs. Variables associated significantly with mortality at univariate analysis were included in the multivariate analysis with a stepwise backward elimination method. The incidence of other clinical outcomes was estimated using the Kaplan–Meier method, compared with a univariate Cox proportional hazards model, and results were reported as the HR and 95% CI. Patients who died before the end of follow-up evaluation were censored at the time of death because they did not develop the outcome. The statistical analysis was performed using the SPSS statistical package (version 24.0 SPSS, Inc. Chicago, IL) and SAS (version 9.4; SAS Institute Inc., Cary, NC).

      Results

       Study Population

      During the study period, 509 patients hospitalized for cirrhosis complications were screened. A total of 349 patients were excluded because of several reasons (Supplementary Figure 1). Thus, 160 patients were included in the study. The mean age was 58 ± 11 years and 70% of patients were male (Table 1). The most common etiology of cirrhosis was alcohol consumption, followed by HCV and HBV infection. Most patients had ascites, a third had hepatic encephalopathy, and a quarter had ACLF. Almost one half of patients had bacterial infections at inclusion in the study. All patients were admitted in a regular ward.
      Table 1Demographic, Clinical, and Laboratory Characteristics of the Whole Population and Patients With or Without Relative Adrenal Insufficiency at Baseline
      VariableWhole population (N = 160)No RAI (N = 82)RAI (N = 78)P value
      Age, y, mean (SD)58 (11)61 (11)57 (10).020
      Male sex, n (%)112 (70)52 (46)60 (54).092
      Etiology of cirrhosis, n (%)
       Alcohol58 (36)35 (43)23 (30).390
       HCV45 (28)21 (26)24 (31)
       HCV + alcohol19 (12)8 (10)11 (14)
       HBV15 (9)7 (9)8 (10)
       Other23 (14)11 (13)12 (15)
      MAP, mm Hg, means (SD)84 (9)85 (9)84 (10).628
      MAP, ≤82 mm Hg, n (%)71 (44)35 (43)36 (46).659
      Heart rate, bpm, means (SD)75 (15)75 (13)76 (17).904
      Ascites, n (%)141 (88)73 (89)68 (87).908
      Hepatic encephalopathy, n (%)49 (31)25 (31)24 (31)1.000
      ACLF, n (%)37 (23)17 (21)20 (26).462
      ACLF grade, n (%)
       Grade 127 (17)14 (17)13 (17).505
       Grade 29 (6)3 (4)6 (8)
       Grade 31 (1)0 (0)1 (1)
      MELD score, means (SD)18 (7)18 (7)19 (7).102
      Child–Pugh score, means (SD)9.1 (1.8)8.9 (1.7)9.3 (1.8).175
      Bacterial infections, n (%)67 (42)29 (35)38 (49).087
      Sepsis, n (%)33 (21)15 (18)18 (23).454
      INR, median (IQR)1.5 (1.3–1.8)1.4 (1.3–1.6)1.6 (1.3–2.0).011
      Bilirubin, mg/dL, median (IQR)3.6 (1.7–7.7)3.0 (1.6–5.9)4.2 (1.7–10.7).154
      Albumin, g/dL, median (IQR)3.1 (2.8–3.5)3.2 (0.6)3.1 (0.6).798
      SCr, mg/dL, median (IQR)1.0 (0.7–1.4)1.0 (0.7–1.5)1.0 (0.7–1.3).355
      Sodium, mmol/L, mean (SD)135 (5)136 (5)135 (5).136
      Hyponatremia, n (%)20 (13)8 (10)12 (15).282
      Leukocytes, ×109/L, median (IQR)4.7 (3.4–6.8)4.5 (2.9–6.3)5.2 (3.6–8.7).023
      CRP, mg/L, median (IQR)14 (6–29)12 (5–23)15 (7–36).085
      ACLF, acute-on-chronic liver failure; CRP, C-reactive protein; HBV, hepatitis B virus, HCV, hepatitis C virus; INR, international normalized ratio; IQR, interquartile range; MAP, mean arterial pressure, MELD, model for end-stage liver disease; RAI, relative adrenal insufficiency; SCr, serum creatinine.

       Prevalence and Clinical Characteristics of Patients With Relative Adrenal Insufficiency

      RAI was found in 78 patients (49%). In Table 1, a comparison of clinical characteristics of patients with or without RAI is reported. Patients with RAI were younger than those without RAI. No differences were found in the etiology and severity of liver disease assessed by MELD and Child–Pugh score. The prevalence of complications of cirrhosis such as ascites, hepatic encephalopathy, and ACLF was similar between the 2 groups, although there was a nonsignificant trend toward a higher prevalence of bacterial infections in patients with RAI than in those without. Patients with RAI showed a higher degree of systemic inflammation than those without RAI, as shown by the higher leukocyte count (5.2 vs 4.5 × 109/L; P = .023) and a nonsignificant trend toward higher levels of CRP.

       Factors Potentially Involved in the Pathogenesis of Relative Adrenal Insufficiency: Pituitary-Adrenal Axis, Substrates for Steroidogenesis, and Inflammatory Cytokines

      We subsequently explored factors potentially involved in the pathogenesis of RAI such as the pituitary-adrenal axis, substrates for steroidogenesis, and inflammatory cytokines (Table 2). Although no difference was found in baseline serum cortisol levels between patients with or without RAI, the former showed higher levels of ACTH than the latter (19 vs 15 ng/L; P = .051).
      Table 2Pituitary Adrenal Axis, Salivary Cortisol, Substrates for Steroidogenesis and Inflammatory Cytokines in Patients With or Without Relative Adrenal Insufficiency
      VariableNo RAI (N = 82)RAI (N = 78)P value
      Pituitary, adrenal axis
       ACTH, ng/L15 (10–31)19 (12–28).051
       Basal cortisol, μg/dL14 (10–19)13 (9–18).449
       Post-ACTH cortisol, μg/dL27 (21–33)19 (15–25)<.001
       Delta cortisol, μg/dL12 (10–15)7 (5–8)<.001
      Salivary cortisol findings
       Basal salivary cortisol, ng/mL2.6 (1.5–4.1)2.4 (1.4–5.1).896
       Post-ACTH salivary cortisol, ng/mL10.2 (5.7–16.4)7.7 (3.3–12.5).007
       Delta salivary cortisol, ng/mL7.2 (3.5–17.6)4.4 (1.7–7.2).002
       Basal salivary cortisone, ng/mL14 (9–21)12 (8–18).140
       Post-ACTH salivary cortisone, ng/mL30 (18–40)21 (13–34).017
       Delta salivary cortisone, ng/mL14 (6–21)9 (3–16).035
       Baseline cortisol/cortisone ratio0.19 (0.14–0.27)0.26 (0.17–0.34).016
       Post-ACTH cortisol/cortisone ratio0.36 (0.28–0.52)0.37 (0.24–0.46).218
      Substrates for steroidogenesis
       Total cholesterol, mmol/L2.00 (1.4–2.7)1.5 (1.2–2.4).005
       HDL cholesterol, mmol/L0.6 (0.3–0.8)0.4 (0.2–0.7).013
       Apolipoprotein A1, g/L0.6 (0.5–0.9)0.5 (0.2–0.8).004
      Inflammatory cytokines
       Tumor necrosis factor α, pg/mL31 (20–49)23 (18–38).146
       Interleukin 1β, pg/mL1.7 (0.5–4.1)1.5 (0.5–3.6).722
       Interleukin 6, pg/mL26 (17–45)38 (20–84).122
       Interleukin 10, pg/mL15 (7–24)20 (6–30).602
      NOTE. Variables are reported as median and interquartile range.
      ACTH, adrenocorticotropic hormone; HDL, high-density lipoprotein; RAI, relative adrenal insufficiency.
      Among substrates involved in steroidogenesis, both total cholesterol and HDL cholesterol were found to be significantly lower in patients with RAI than in those without. Furthermore, apolipoprotein-A1 was significantly lower in patients with RAI than in those without. Inflammatory cytokines were not significantly different between patients with RAI and those without. In multivariate analysis, age (OR, 0.95; 95% CI, 0.91–0.99; P = .030), HDL cholesterol level (OR, 0.30; 95% CI, 0.09–0.94; P = .039), and leukocyte count (OR, 3.10; 95% CI, 1.39–6.9; P = .006) were found to be independent predictors of RAI (Supplementary Table 1). In patients without infections, HDL cholesterol was the only predictor of RAI (OR, 0.28; 95% CI, 0.08–0.95; P = .041), whereas in patients with infections, age (OR, 0.94; 95% CI, 0.88–1.00; P = .048) and leukocytes (OR, 3.54; 95% CI, 1.06–11.89; P = .041) were independent predictors of RAI.

       Salivary Cortisol Findings

      Salivary cortisol was evaluated before and after SST in 121 patients. It was not possible to analyze salivary samples in 39 patients because of insufficient quantity of saliva (N = 18), contamination of saliva with blood (N = 10), and evidence of blood in the mouth on the day of SST (N = 11). No differences were found in baseline salivary cortisol levels between patients with or without RAI, whereas post-ACTH cortisol and delta salivary cortisol were significantly lower in patients with RAI than in those without (Table 2). A significant direct correlation was found between serum and salivary cortisol considering both baseline levels (R = 0.474; P < .001) (Figure 1A) and the delta values (R = 0.391; P < .001) (Figure 1B). The analysis of salivary cortisone showed similar findings (data not shown).
      Figure thumbnail gr1
      Figure 1(A) Correlation between salivary cortisol and serum cortisol at baseline and the (B) correlation between the delta salivary cortisol and delta serum cortisol after the Short Synacthen Test.

       Mortality

      Patients were followed up for a mean of 68 ± 32 days. During the 90-day follow-up period, 28 patients died (18%), 30 underwent a transplant (19%), 3 were lost to follow-up evaluation (2%), and 99 survived (62%). As expected, patients who died had more advanced liver disease and a higher prevalence of hepatic encephalopathy, ACLF, and bacterial infections than those who survived (Supplementary Table 2). Moreover, nonsurvivors showed a poorer hemodynamic status and renal function than survivors. Leukocytes and CRP levels were significantly higher in patients who died than in those who survived. Finally, nonsurvivors had a significantly higher prevalence of RAI (71% vs 42%; sHR, 2.92; P = .009) and lower values of delta cortisol (8 vs 10 μg/dL; sHR, 0.67; P = .037) than those who survived. The cumulative incidence of 90-day mortality was 26% in patients with RAI and 10% in those without (P = .008) (Figure 2A). Interestingly, even in the subgroup of patients with ACLF or bacterial infections, the cumulative incidence of 90-day mortality was significantly higher in patients with RAI than in those without (40% vs 12%; P = .042; and 38% vs 14%; P = .031, respectively). In multivariate analysis, age (sHR, 1.07; P = .004), model for end-stage liver disease score (sHR, 1.1; P = .014), serum sodium level (sHR, 0.91; P = .003), bacterial infections (sHR, 3.08; P = .011), hepatic encephalopathy (sHR, 2.21; P = .041), and RAI (sHR, 4.83; P = .001) were found to be independent predictors of 90-day mortality (Table 3). The cumulative incidence of 28-day mortality was significantly higher in patients with RAI than in those without (18% vs 4%; P = .003), and RAI was an independent predictor of 28-day mortality (sHR, 4.12; P = .031) (Supplementary Table 3).
      Figure thumbnail gr2
      Figure 2(A) Cumulative incidence of 90-day mortality in patients with or without relative adrenal insufficiency (RAI) and the (B) cumulative incidence of 28-day mortality in patients with the European Association for the Study of the Liver-Chronic Liver Failure (EASL-CLIF) definition of acute-on-chronic liver failure (ACLF), patients without ACLF and patients without ACLF but with RAI and renal dysfunction, or mild/moderate hepatic encephalopathy (RAI ACLF). *P < .001 vs RAI ACLF; #P = .001 vs EASL-CLIF ACLF; and §P = .582 vs EASL-CLIF ACLF.
      Table 3Independent Predictors of 90-Day Mortality in Patients With Acute Decompensation of Cirrhosis
      VariablesHR95% CIP value
      Age1.071.02–1.11.004
      Bacterial infections3.081.30–7.31.011
      Hepatic encephalopathy2.211.03–4.74.041
      MELD1.101.02–1.19.014
      Serum sodium0.910.86–0.97.003
      RAI4.832.00–11.68.001
      MELD, model for end stage liver disease; RAI, relative adrenal insufficiency; sHR, subdistribution hazard ratio.

       Clinical Outcomes in Patients With or Without Relative Adrenal Insufficiency

      During the follow-up period, patients with RAI showed a significantly higher incidence of bacterial infections (62% vs 45%; P = .038), sepsis (45% vs 17%; P = .001), and septic shock (12% vs 3%; P = .038) than those without RAI (Table 4). Patients with RAI had a trend toward a significantly higher incidence of AKI (41% vs 26%; P = .051) and hepatic encephalopathy (44 vs 29; P = .053) than those without. Finally, patients with RAI showed a significantly higher incidence of new organ failures (33% vs 14%; P = .014) and ACLF (32% vs 16%; P = .037) than those without.
      Table 4Outcomes That Occurred During the 90-Day Follow-Up Period According to the Presence of RAI
      OutcomesNo RAIRAIHR (95% CI)P value
      Development of hyponatremia, n (%)
      Hyponatremia was defined as a serum sodium level less than 130 mmol/L. Only patients without hyponatremia at inclusion (n = 140) were included in this analysis.
      17 (25)27 (46)1.83 (1.11–3.03).018
      Development of AKI, n (%)20 (26)29 (41)1.75 (1.00–3.08).051
      Development of HE, n (%)22 (29)29 (44)1.72 (0.99–2.97).053
      Development of new organ failures, n (%)
      Composite outcome of liver failure, renal failure, coagulation failure, brain failure, circulatory failure, and lung failure.
      11 (14)22 (33)2.45 (1.20–5.02).014
      Development of ACLF, n (%)
      Only patients without ACLF at inclusion (n = 123) were included in this analysis.
      10 (16)18 (32)2.27 (1.05–4.88).037
      Development of bacterial infections, n (%)
      New bacterial infections (second infections for patients already infected at enrollment, first infection for those not infected at enrollment).
      34 (45)43 (62)1.60 (1.03–2.50).038
      Development of sepsis, n (%)13 (17)30 (45)2.95 (1.54–5.67).001
      Development of septic shock, n (%)2 (3)8 (12)4.94 (1.10–22.24).038
      NOTE. Percentages refer to the 90-day probability of the outcome; the HR was estimated with a univariate Cox proportional hazards method.
      ACLF, acute-on-chronic liver failure; AKI, acute kidney injury; HE, hepatic encephalopathy; HR, hazard ratio; RAI, relative adrenal insufficiency.
      a Hyponatremia was defined as a serum sodium level less than 130 mmol/L. Only patients without hyponatremia at inclusion (n = 140) were included in this analysis.
      b Composite outcome of liver failure, renal failure, coagulation failure, brain failure, circulatory failure, and lung failure.
      c Only patients without ACLF at inclusion (n = 123) were included in this analysis.
      d New bacterial infections (second infections for patients already infected at enrollment, first infection for those not infected at enrollment).

       Relative Adrenal Insufficiency: A New Organ Failure in the Assessment of Acute-on-Chronic Liver Failure?

      According to the aforementioned results, we explored the hypothesis that RAI can be used as a new organ failure in the evaluation of ACLF. According to the EASL-CLIF consortium definition, ACLF is a syndrome characterized by organ failures and a 28-day mortality rate greater than 15%.
      • Moreau R.
      • Jalan R.
      • Gines P.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      Grade 1 ACLF is defined by the occurrence of a single organ failure with a 28-day mortality rate greater than 15%. Renal failure is the only single organ failure associated with 28-day mortality exceeding 15%, while all the other organ failures (liver, brain, coagulation, circulation, and lung) need to be associated with a renal dysfunction (serum creatinine concentration, 1.5–1.9 mg/dL) and/or a mild/moderate hepatic encephalopathy (grades 1–2 according to West-Haven criteria) to exceed the 15% threshold. Because patients with RAI and without ACLF had a cumulative incidence of 28-day mortality of 10%, we explored 28-day mortality in patients without ACLF, with RAI, plus a mild renal dysfunction and/or a mild/moderate hepatic encephalopathy. The cumulative incidence of 28-day mortality in these patients was 25%, similar to that observed in patients with ACLF grade 1 (22%) (Supplementary Figure 2). Thus, considering RAI as a new nonrenal organ failure, 18 of 123 patients without ACLF (15%) according to the EASL-CLIF definition were reclassified as having ACLF. The latter had a cumulative incidence of 28-day mortality of 28% that was similar to those observed in patients with EASL-CLIF ACLF (22%; P = NS), and significantly higher than that of the remaining patients without ACLF (5%; P < .001) (Figure 2B). With this approach, the prevalence of ACLF in the whole population was increased by 11% (Supplementary Figure 3A). As far as the grading of ACLF, there was a decrease in ACLF grade 1 episodes (55% vs 73% of ACLF) and an increase in ACLF grades 2 (33% vs 22%) and 3 episodes (13% vs 3%) (Supplementary Figure 3B).

      Discussion

      This was a large study investigating the prevalence and the clinical impact of RAI in patients with cirrhosis. The results of our study showed some relevant findings. We did not use a baseline threshold of serum total cortisol to define RAI in our cohort, but we preferred dynamic changes of serum total cortisol after SST.
      • Angeli P.
      • Bernardi M.
      • Villanueva C.
      • et al.
      EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis.
      Indeed, it has been shown clearly that criteria based on absolute values of serum total cortisol overestimate RAI, mainly because of inaccurate determinations related to hypoalbuminemia and low levels of cortisol binding globulin.
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      ,
      • Tan T.
      • Chang L.
      • Woodward A.
      • et al.
      Characterising adrenal function using directly measured plasma free cortisol in stable severe liver disease.
      ,
      • Thevenot T.
      • Borot S.
      • Remy-Martin A.
      • et al.
      Assessment of adrenal function in cirrhotic patients using concentration of serum-free and salivary cortisol.
      Nevertheless, we also assessed dynamic changes of cortisol and cortisone in the saliva of our patients because they closely reflect serum free cortisol concentrations.
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      ,
      • Thevenot T.
      • Borot S.
      • Remy-Martin A.
      • et al.
      Assessment of adrenal function in cirrhotic patients using concentration of serum-free and salivary cortisol.
      Thus, salivary cortisol was used as a surrogate of serum free cortisol, the biologically active fraction of cortisol, which cannot be assessed routinely.
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      ,
      • Thevenot T.
      • Borot S.
      • Remy-Martin A.
      • et al.
      Assessment of adrenal function in cirrhotic patients using concentration of serum-free and salivary cortisol.
      Salivary cortisone was used to measure the amount of free cortisol metabolized to cortisone by the 11β-hydroxysteroid dehydrogenase type 2 in the salivary glands.
      • Perogamvros I.
      • Keevil B.G.
      • Ray D.W.
      • et al.
      Salivary cortisone is a potential biomarker for serum free cortisol.
      The delta serum total cortisol after SST had a strong correlation with the delta salivary cortisol and cortisone, proving that the change of serum total cortisol after SST is a reliable measure of RAI in patients with cirrhosis. Thus, the first relevant finding of our study was to confirm that the change of serum total cortisol after SST is a reliable measure of RAI in patients with cirrhosis. Conversely, the assessment of salivary cortisol had a technical failure in 24% of patients, thus, it is unfeasible for the routine evaluation of RAI in cirrhosis. RAI was diagnosed in 49% of patients with an acute decompensation of cirrhosis. The prevalence of RAI found in our cohort was higher than that reported by Acevedo et al,
      • Acevedo J.
      • Fernández J.
      • Prado V.
      • et al.
      Relative adrenal insufficiency in decompensated cirrhosis: relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death.
      despite similar diagnostic criteria. This discrepancy can be explained by the lower level of serum cholesterol observed in our patients.
      This observation leads us to introduce the second relevant finding of our study that concerns the pathogenesis of RAI. We proved that low levels of HDL cholesterol were associated independently with RAI. Thus, we confirmed definitively that HDL cholesterol could be responsible for a deficit of substrates for steroidogenesis in patients with cirrhosis.
      • Tsai M.-H.
      • Huang H.-C.
      • Peng Y.-S.
      • et al.
      Critical illness–related corticosteroid insufficiency in cirrhotic patients with acute gastroesophageal variceal bleeding: risk factors and association with outcome*.
      ,
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      Nevertheless, the pathogenesis of RAI in patients with cirrhosis is probably much more complex. In fact, we showed that a rough surrogate marker of inflammation, such as the blood leukocyte count, was an independent predictor of RAI. Indeed, immune cells and inflammatory cytokines can inhibit steroidogenesis, competing with ACTH at the receptor level or toward a direct effect on adrenal glands.
      • Cooper M.S.
      • Stewart P.M.
      Corticosteroid insufficiency in acutely ill patients.
      Therefore, a potential cause–effect relationship between inflammation and RAI may exist in patients with cirrhosis. However, RAI may contribute per se to excessive inflammation and/or a reduced tolerance to inflammation.
      • Cooper M.S.
      • Stewart P.M.
      Corticosteroid insufficiency in acutely ill patients.
      Given the nature of our study, we cannot support 1 of these 2 interpretations, which, however, are not mutually exclusive.
      This pathophysiological finding is the background of the most relevant clinical finding of our study. In fact, it has been shown that ACLF is associated with excessive inflammation or a reduced tolerance to systemic inflammation,
      • Clària J.
      • Stauber R.E.
      • Coenraad M.J.
      • et al.
      Systemic inflammation in decompensated cirrhosis: characterization and role in acute-on-chronic liver failure.
      ,
      • Moreau R.
      • Jalan R.
      • Gines P.
      • et al.
      Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
      thus, it can be hypothesized that the inadequate function of adrenal glands could disturb the host physiology, affecting the ability to control/shutdown the inflammatory response and/or the tolerance capacity of the single organs. Accordingly, we explored the possibility that RAI could be considered to better diagnose and classify patients with ACLF. RAI was associated with an increase in the cumulative incidence of mortality in all patients with acute decompensation of cirrhosis, as well as in patients who developed an ACLF, suggesting that RAI is capable of disturbing the homeostasis that is crucial for the host tolerance to inflammation.
      • Medzhitov R.
      • Schneider D.S.
      • Soares M.P.
      Disease tolerance as a defense strategy.
      Thus, by focusing on RAI as a potential additional organ failure to the EASL-CLIF consortium classification, we observed that in patients with decompensated cirrhosis and RAI, but without ACLF, the cumulative incidence of mortality at 28 days was less than 15% (ie, the threshold used for defining ACLF). However, when RAI was associated with renal dysfunction and/or mild to moderate hepatic encephalopathy, the cumulative incidence of mortality rate at 28 days was greater than 15%. Thus, RAI per se showed similar characteristics to the organ failures other than the kidney used in the EASL-CLIF classification. Consequently, when RAI was considered as an additional organ failure in the EASL-CLIF classification of ACLF, an increase in the prevalence as well as a shift of ACLF from grade 1 toward grades 2 and 3 were observed.
      As to the other clinical outcomes of RAI in patients with decompensated cirrhosis, we confirmed that it was an independent predictor of 90-day mortality and it was associated with a higher risk of developing an infection, sepsis, and septic shock. These results are well in keeping with those of Acevedo et al,
      • Acevedo J.
      • Fernández J.
      • Prado V.
      • et al.
      Relative adrenal insufficiency in decompensated cirrhosis: relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death.
      but not with those of other investigators,
      • Galbois A.
      • Rudler M.
      • Massard J.
      • et al.
      Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
      • Tan T.
      • Chang L.
      • Woodward A.
      • et al.
      Characterising adrenal function using directly measured plasma free cortisol in stable severe liver disease.
      • Fede G.
      • Spadaro L.
      • Tomaselli T.
      • et al.
      Assessment of adrenocortical reserve in stable patients with cirrhosis.
      ,
      • Risso A.
      • Alessandria C.
      • Mezzabotta L.
      • et al.
      Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: relationship and effect on survival.
      ,
      • Fede G.
      • Spadaro L.
      • Tomaselli T.
      • et al.
      Comparison of total cortisol, free cortisol, and surrogate markers of free cortisol in diagnosis of adrenal insufficiency in patients with stable cirrhosis.
      probably because of the different criteria used to define RAI and/or because of the smaller sample sizes.
      We realize that our study had some limitations. Unfortunately, we did not assess serum free cortisol, which would have enforced our findings further. Furthermore, we did not repeat SST during the follow-up evaluation, thus we were not able to show if RAI persists or resolves after the recovery from acute decompensation. This was a single-center study and although results are in keeping with other investigators they need further external validation. Finally, we did not treat patients with RAI with steroids and therefore we lack the confirmation that the treatment can improve clinical outcomes in these patients. Future studies should address this issue.
      In conclusion, we showed that RAI is common in patients with an acute decompensation of cirrhosis. The deficit of substrates for steroidogenesis and the degree of inflammatory response seems to be associated with the severity of RAI. RAI is associated with the development of bacterial infections, sepsis, septic shock, ACLF, and short-term mortality. Finally, considering RAI as additional organ failure in the EASL-CLIF classification, it has an impact on short-term mortality similar to that of the other nonrenal organ failures. Thus, RAI should be considered in clinical practice for better stratification of patients with ACLF.

      Supplementary Methods

       Assessment of Salivary Cortisol and Cortisone

      Salivary cortisol and cortisone levels were measured by a homemade liquid chromatography–tandem mass spectrometry, with automated sample preparation. Briefly, the sample was kept frozen from collection time until it was analyzed. On the day of analysis, the samples were thawed and centrifuged at 3000 rpm for 10 minutes at room temperature. Solid phase extraction on a μ-elution 96-well plate was performed by a liquid handling platform. After extraction, the eluates underwent liquid chromatography–tandem mass spectrometry. Salivary cortisol and cortisone were measured using an Agilent high-performance liquid chromatography series 1200 triple-quadrupole mass spectrometer Agilent 6430 equipped with an electrospray ionization source in positive ionization mode (Agilent Technologies, Palo Alto, CA). Quantitative analysis was performed in the multiple-reaction monitoring mode. This method was validated according to the Medical laboratories - Requirements for quality and competence ISO15189:2012. The method was linear up to 55.4 nmol/L (20.1 ng/mL) and 51 nmol/L (18.4 ng/mL), with a lower quantification limit of 0.51 nmol/L (0.2 ng/mL) and 0.55 nmol/L (0.2 ng/mL) for cortisol and cortisone, respectively. Within- and between-run coefficients of variation were less than 10% for both steroids and mean recoveries were 106%. Salivary cortisol and cortisone levels higher than the upper limit of linearity were retested with appropriate dilution.

       Assessment of Acute-on-Chronic Liver Failure and Sepsis

      Because the EASL-CLIF definition of ACLF was published after the initiation of the study (in 2013), it was assessed using the available data for 62 patients. Indeed, in the study protocol, the collection of variables needed for the assessment of ACLF (bilirubin, international normalized ratio, platelet count, serum creatinine, West Haven classification of hepatic encephalopathy, mean arterial pressure, and oxygen saturation) already was planned, both at admission and during the hospitalization. For the remaining 98 patients, the assessment of ACLF was performed prospectively after an amendment to the study protocol.
      As far as the sepsis definition is concerned, assessment of sepsis according to Sepsis-3 criteria was performed retrospectively because these criteria were published in 2016. Data used for the assessment of Sepsis-3 criteria (change in Sequential Organ Failure Assessment score ≥2 in a patient with a new infection) already were available in the database. The preadmission Sequential Organ Failure Assessment score was assessed retrospectively for the assessment of sepsis at baseline .
      Figure thumbnail fx1
      Supplementary Figure 1Flow chart of patients included in the study. HCC, hepatocellular carcinoma; HIV, human immunodeficiency virus; RAI, relative adrenal insufficiency.
      Figure thumbnail fx2
      Supplementary Figure 2Cumulative incidence of 28-day mortality in patients with an European Foundation– Chronic Liver Failure (EF-CLIF) definition of grade 1 acute-on-chronic liver failure (EF-CLIF ACLF) and in patients without ACLF but with relative adrenal insufficiency and renal dysfunction, or mild to moderate hepatic encephalopathy (RAI ACLF). The comparison was performed using the Gray test. RAI, relative adrenal insufficiency.
      Figure thumbnail fx3
      Supplementary Figure 3(A) Prevalence of acute-on-chronic liver failure and (B) acute on-chronic liver failure grade according to the European Association for the Study of the Liver-Chronic Liver Failure criteria (EASL-CLIF) and considering relative adrenal insufficiency as the other nonrenal organ failure for defining ACLF (RAI ACLF). ACLF, acute-on-chronic liver failure; RAI, relative adrenal insufficiency.
      Supplementary Table 1Independent Predictors of Relative Adrenal Insufficiency in Patients With Acute Decompensation of Cirrhosis
      VariableOR95% CIP value
      Age, y0.950.91–0.99.030
      HDL cholesterol, mg/dL0.300.09–0.94.039
      Leukocytes, ×109/L3.101.39–6.9.006
      HDL, high-density lipoprotein; OR, odds ratio.
      Supplementary Table 2Risk Factors for 90-Day Mortality in Patients With an Acute Decompensation of Cirrhosis
      VariableSurvivors (N = 99)Nonsurvivors (N = 28)sHRP value
      Age, y, mean (SD)59 (12)62 (9)1.03 (1.00–1.06).080
      Male sex, n (%)63 (64)22 (79)1.64 (0.67–4.01).283
      Etiology of cirrhosis (HCV), n (%)25 (25)9 (32)1.24 (0.56–2.72).594
      MAP, mm Hg, mean (SD)85 (10)81 (7)0.97 (0.94–1.00).031
      Heart rate, bpm, mean (SD)74 (13)80 (21)1.03 (0.99–1.07).124
      Ascites, n (%)80 (81)28 (100)26.66 (0.36–1976.20).135
      Hepatic encephalopathy, n (%)23 (23)13 (46)2.20 (1.05–4.60).036
      ACLF, n (%)9 (9)10 (36)2.04 (0.95–4.41).069
      MELD score, mean (SD)15 (5)22 (7)1.08 (1.03–1.13).002
      Child–Pugh score, mean (SD)8.6 (1.7)10.1 (1.5)1.40 (1.17–1.69)<.001
      Bacterial infections, n (%)35 (35)18 (64)2.71 (1.25–5.85).011
      INR, median (IQR)1.4 (1.3–1.6)1.6 (1.4–2.0)6.14 (1.18–32.04).031
      Bilirubin, mg/dL, median (IQR)2.5 (1.4–4.8)6.2 (2.9–14.6)1.58 (1.12–2.22).009
      Albumin, g/dL, mean (SD)3.2 (0.7)3.0 (0.6)0.96 (0.90–1.03).252
      SCr, mg/dL, median (IQR)0.9 (0.6–1.2)1.3 (0.9–1.4)1.91 (1.05–3.48).034
      Sodium, mmol/L, mean (SD)136 (4)133 (5)0.91 (0.85–0.97).004
      Leukocytes, ×109/L, median (IQR)4.5 (2.9–6.2)5.5 (3.6–11.4)2.51 (1.68–3.77)<.001
      CRP, mg/L, median (IQR)11 (5–21)18 (9–35)1.35 (0.99–1.83).056
      ACTH, ng/L, median (IQR)17 (11–27)20 (12–29)1.90 (0.99–3.64).054
      Basal cortisol, μg/dL, median (IQR)12 (9–17)16 (10–21)2.26 (0.98–5.17).055
      Delta cortisol, μg/dL, median (IQR)10 (7–12)8 (5–10)0.67 (0.47–0.98).037
      RAI, n (%)42 (42)20 (71)2.92 (1.30–6.56).009
      Basal salivary cortisol, ng/mL, median (IQR)2.2 (1.2–4.1)3.4 (2.3–9.2)1.80 (1.17–2.76).008
      Delta salivary cortisol, ng/mL, median (IQR)5 (3.1–10.2)5 (1.5–10.3)0.85 (0.53–1.38).518
      Basal salivary cortisone, ng/mL, median (IQR)12 (7–18)17 (10–20)1.84 (0.93–3.61).079
      Delta salivary cortisone, ng/mL, median (IQR)11 (5–16)15 (5–27)1.27 (0.66–2.44).470
      NOTE. Liver transplantation was considered a competing event for death.
      ACLF, acute on chronic liver failure; ACTH, adrenocorticotropic hormone; CRP, C-reactive protein; HCV, hepatitis C virus; INR, international normalized ratio; IQR, interquartile range; MAP, mean arterial pressure; MELD, model for end stage liver disease; RAI, relative adrenal insufficiency; SCr, serum creatinine; sHR, subdistribution hazard ratio.
      Supplementary Table 3Independent Predictors of 28-Day Mortality in Patients With Acute Decompensation of Cirrhosis
      VariablesHR95% CIP value
      Bacterial infections2.430.87–6.82.091
      MELD1.111.03–1.20.005
      Serum sodium0.910.84–0.99.032
      RAI4.121.14–14.90.031
      MELD, model for end stage liver disease; RAI, relative adrenal insufficiency; sHR, subdistribution hazard ratio.

      References

        • Cooper M.S.
        • Stewart P.M.
        Corticosteroid insufficiency in acutely ill patients.
        N Engl J Med. 2003; 348: 727-734
        • Marik P.E.
        • Pastores S.M.
        • Annane D.
        • et al.
        Recommendations for the diagnosis and management of corticosteroid insufficiency in critically ill adult patients: consensus statements from an international task force by the American College of Critical Care Medicine.
        Crit Care Med. 2008; 36: 1937-1949
        • Annane D.
        • Sébille V.
        • Troché G.
        • et al.
        A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin.
        JAMA. 2000; 283: 1038-1045
        • Fernández J.
        • Escorsell A.
        • Zabalza M.
        • et al.
        Adrenal insufficiency in patients with cirrhosis and septic shock: effect of treatment with hydrocortisone on survival.
        Hepatology. 2006; 44: 1288-1295
        • Tsai M.-H.
        • Peng Y.-S.
        • Chen Y.-C.
        • et al.
        Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock.
        Hepatology. 2006; 43: 673-681
        • Arabi Y.M.
        • Aljumah A.
        • Dabbagh O.
        • et al.
        Low-dose hydrocortisone in patients with cirrhosis and septic shock: a randomized controlled trial.
        Can Med Assoc J. 2010; 182: 1971-1977
        • Tsai M.-H.
        • Huang H.-C.
        • Peng Y.-S.
        • et al.
        Critical illness–related corticosteroid insufficiency in cirrhotic patients with acute gastroesophageal variceal bleeding: risk factors and association with outcome*.
        Crit Care Med. 2014; 42: 2546-2555
        • Galbois A.
        • Rudler M.
        • Massard J.
        • et al.
        Assessment of adrenal function in cirrhotic patients: salivary cortisol should be preferred.
        J Hepatol. 2010; 52: 839-845
        • Tan T.
        • Chang L.
        • Woodward A.
        • et al.
        Characterising adrenal function using directly measured plasma free cortisol in stable severe liver disease.
        J Hepatol. 2010; 53: 841-848
        • Fede G.
        • Spadaro L.
        • Tomaselli T.
        • et al.
        Assessment of adrenocortical reserve in stable patients with cirrhosis.
        J Hepatol. 2011; 54: 243-250
        • Thevenot T.
        • Borot S.
        • Remy-Martin A.
        • et al.
        Assessment of adrenal function in cirrhotic patients using concentration of serum-free and salivary cortisol.
        Liver Int. 2011; 31: 425-433
        • Acevedo J.
        • Fernández J.
        • Prado V.
        • et al.
        Relative adrenal insufficiency in decompensated cirrhosis: relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death.
        Hepatology. 2013; 58: 1757-1765
        • Graupera I.
        • Pavel O.
        • Hernandez-Gea V.
        • et al.
        Relative adrenal insufficiency in severe acute variceal and non-variceal bleeding: influence on outcomes.
        Liver Int. 2015; 35: 1964-1973
        • Risso A.
        • Alessandria C.
        • Mezzabotta L.
        • et al.
        Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: relationship and effect on survival.
        Dig Liver Dis. 2015; 47: 702-708
        • Fede G.
        • Spadaro L.
        • Tomaselli T.
        • et al.
        Comparison of total cortisol, free cortisol, and surrogate markers of free cortisol in diagnosis of adrenal insufficiency in patients with stable cirrhosis.
        Clin Gastroenterol Hepatol. 2014; 12: 504-512.e8
        • Jang J.Y.
        • Kim T.Y.
        • Sohn J.H.
        • et al.
        Relative adrenal insufficiency in chronic liver disease: its prevalence and effects on long-term mortality.
        Aliment Pharmacol Ther. 2014; 40: 819-826
        • Fede G.
        • Spadaro L.
        • Tomaselli T.
        • et al.
        Adrenocortical dysfunction in liver disease: a systematic review.
        Hepatology. 2012; 55: 1282-1291
        • Clària J.
        • Stauber R.E.
        • Coenraad M.J.
        • et al.
        Systemic inflammation in decompensated cirrhosis: characterization and role in acute-on-chronic liver failure.
        Hepatology. 2016; 64: 1249-1264
        • Bernardi M.
        • Moreau R.
        • Angeli P.
        • et al.
        Mechanisms of decompensation and organ failure in cirrhosis: from peripheral arterial vasodilation to systemic inflammation hypothesis.
        J Hepatol. 2015; 63: 1272-1284
        • Moreau R.
        • Jalan R.
        • Gines P.
        • et al.
        Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis.
        Gastroenterology. 2013; 144: 1426-1437.e9
        • Antonelli G.
        • Ceccato F.
        • Artusi C.
        • et al.
        Salivary cortisol and cortisone by LC–MS/MS: validation, reference intervals and diagnostic accuracy in Cushing’s syndrome.
        Clin Chim Acta. 2015; 451: 247-251
        • Angeli P.
        • Bernardi M.
        • Villanueva C.
        • et al.
        EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis.
        J Hepatol. 2018; 69: 406-460
        • Angeli P.
        • Gines P.
        • Wong F.
        • et al.
        Diagnosis and management of acute kidney injury in patients with cirrhosis: revised consensus recommendations of the International Club of Ascites.
        J Hepatol. 2015; 62: 968-974
        • Piano S.
        • Bartoletti M.
        • Tonon M.
        • et al.
        Assessment of Sepsis-3 criteria and quick SOFA in patients with cirrhosis and bacterial infections.
        Gut. 2018; 67: 1892-1899
        • Fine J.P.
        • Gray R.J.
        A proportional hazards model for the subdistribution of a competing risk.
        J Am Stat Assoc. 1999; 94: 496-509
        • Perogamvros I.
        • Keevil B.G.
        • Ray D.W.
        • et al.
        Salivary cortisone is a potential biomarker for serum free cortisol.
        J Clin Endocrinol Metab. 2010; 95: 4951-4958
        • Medzhitov R.
        • Schneider D.S.
        • Soares M.P.
        Disease tolerance as a defense strategy.
        Science. 2012; 335: 936-941

      Linked Article

      • Adrenal Insufficiency in Cirrhosis: Don’t Forget the Hypothalamus
        Clinical Gastroenterology and HepatologyVol. 19Issue 5
        • Preview
          We read with interest the recent article reporting a very high rate of relative adrenal insufficiency (RAI) in patients with cirrhosis admitted to the hospital with acute decompensation.1 We would like to add to their hypothesis that adrenal insufficiency in cirrhosis is caused by adrenal hypoperfusion and impaired substrates for steroidogenesis evidenced by reduced total cholesterol, high-density lipoprotein, and apolipoprotein A1 in patients with RAI. Although RAI may be multifactorial, we previously showed that adrenal insufficiency in patients with advanced cirrhosis predominantly is the result of a defect in the hypothalamic-pituitary-adrenal axis as opposed to an intrinsic adrenal gland defect.
        • Full-Text
        • PDF