Potential conflict of interest: Nothing to report.
Liver Failure/Cirrhosis/Portal Hypertension
Relative adrenal insufficiency in decompensated cirrhosis: Relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death
Article first published online: 7 AUG 2013
© 2013 by the American Association for the Study of Liver Diseases
Volume 58, Issue 5, pages 1757–1765, November 2013
How to Cite
Acevedo, J., Fernández, J., Prado, V., Silva, A., Castro, M., Pavesi, M., Roca, D., Jimenez, W., Ginès, P. and Arroyo, V. (2013), Relative adrenal insufficiency in decompensated cirrhosis: Relationship to short-term risk of severe sepsis, hepatorenal syndrome, and death. Hepatology, 58: 1757–1765. doi: 10.1002/hep.26535
Supported by FIS PI10/01373; Juan Acevedo was supported by a grant from Instituto de Salud Carlos III (CM08/00129) and Hospital Clinic. CIBEREHD is funded by the Instituto de Salud Carlos III.
- Issue published online: 30 OCT 2013
- Article first published online: 7 AUG 2013
- Accepted manuscript online: 31 MAY 2013 08:39AM EST
- Manuscript Accepted: 15 MAY 2013
- Manuscript Revised: 9 MAY 2013
- Manuscript Received: 19 FEB 2013
The prevalence of relative adrenal insufficiency (RAI) in critically ill cirrhosis patients with severe sepsis is over 60% and associated features include poor liver function, renal failure, refractory shock, and high mortality. RAI may also develop in noncritically ill cirrhosis patients but its relationship to the clinical course has not yet been assessed. The current study was performed in 143 noncritically ill cirrhosis patients admitted for acute decompensation. Within 24 hours after hospitalization adrenal function, plasma renin activity, plasma noradrenaline and vasopressin concentration, and serum levels of nitric oxide, interleukin-6 and tumor necrosis factor alpha were determined. RAI was defined as a serum total cortisol increase <9 μg/dL after 250 μg of intravenous corticotropin from basal values <35 μg/dL. Patients were followed for 3 months. RAI was detected in 26% of patients (n = 37). At baseline, patients with RAI presented with lower mean arterial pressure (76 ± 12 versus 83 ± 14 mmHg, P = 0.009) and serum sodium (131 ± 7 versus 135 ± 5 mEq/L, P = 0.007) and higher blood urea nitrogen (32 ± 24 versus 24 ± 15 mg/dl, P = 0.06), plasma renin activity (7.1 ± 9.9 versus 3.4 ± 5.6 ng/mL*h, P = 0.03), and noradrenaline concentration (544 ± 334 versus 402 ± 316 pg/mL, P = 0.02). During follow-up, patients with RAI exhibited a higher probability of infection (41% versus 21%, P = 0.008), severe sepsis (27% versus 9%, P = 0.003), type-1 hepatorenal syndrome (HRS) (16% versus 3%, P = 0.002), and death (22% versus 7%, P = 0.01). Conclusion: RAI is frequent in noncritically ill patients with acute decompensation of cirrhosis. As compared with those with normal adrenal function, patients with RAI have greater impairment of circulatory and renal function, higher probability of severe sepsis and type-1 HRS, and higher short-term mortality. (Hepatology 2013;58:1757–1765)
acute-on-chronic liver failure
Barcelona-Clinic Liver Cancer
blood urea nitrogen
calculated free cortisol
free cortisol index
global initiative for chronic obstructive lung disease
hepatitis C virus
high density lipoprotein
human immunodeficiency virus
International Ascites Club
intensive care unit
low density lipoprotein
New York Heart Association
model for endstage liver disease
relative adrenal insufficiency
systemic inflammatory response syndrome
short Synacthen test
tumor necrosis factor alpha;
Relative adrenal insufficiency (RAI) is a syndrome characterized by an inadequate production of cortisol with respect to peripheral demands.[1, 2] It was first described and has been mainly studied in critically ill patients (severe sepsis, septic shock, head injury, pancreatitis, burns, and major surgery) who require high circulating cortisol levels to modulate systemic inflammatory response, maintain the vascular tone and permeability, and adapt metabolism to stress.[1-7] During the last decade evidence has been presented that RAI may also be an important feature in patients with cirrhosis.[8-22]
The first studies on RAI in critically ill patients with cirrhosis were published between 2003 and 2008 and included mainly patients with decompensated cirrhosis and severe sepsis or septic shock. These studies showed a very high prevalence of RAI (51%-77%) and a clear association with poor liver function, renal failure, refractory shock, and hospital mortality.[8-11] Two recent studies confirm the high prevalence of RAI in patients with cirrhosis and septic shock (76%) or gastrointestinal hemorrhage (60%).
Recent but limited data suggest that RAI can also occur in noncritically ill cirrhosis patients.[14-22] The reported prevalence of this entity, which is now refer to as “hepatoadrenal syndrome,”[9, 22] ranges in the different studies between 7% and 49%, depending on the methodology used for RAI diagnosis.[15-22] However, there are no data on the relationship between RAI and clinical course of noncritically ill cirrhosis patients.
This article reports the results of a prospective study evaluating adrenal function in a large series of patients admitted to the hospital for the treatment of acute decompensation of cirrhosis. Patients were followed for 3 months. The aim of the study was to assess if RAI is associated with differences in the natural course of the disease and survival.
Patients and Methods
In all, 143 patients admitted to our Liver Unit between 2008 and 2010 for the treatment of an acute decompensation of cirrhosis (ascites, hepatic encephalopathy, bacterial infection, variceal bleeding, and hepatorenal syndrome [HRS]) were studied prospectively. Diagnosis of cirrhosis was established by histology or by clinical, analytical, and ultrasonographic findings. Inclusion criteria were age between 18 and 80 years and hospitalization due to clinical decompensation of cirrhosis. Exclusion criteria were: human immunodeficiency virus (HIV) infection, previous transplantation or any other type of immunodeficiency, steroid treatment, pituitary or adrenal disease, advanced hepatocellular carcinoma (Barcelona-Clinic Liver Cancer [BCLC] stage B, C, or D), severe chronic heart (New York Heart Association [NYHA] class III or IV) or pulmonary disease (global initiative for chronic obstructive lung disease [GOLD] III or IV), chronic hemodialysis, time between hospital admission and baseline evaluation >24 hours, severe sepsis, hypovolemic or septic shock, acute respiratory distress syndrome, and refusal of patient to participate. Patients or their relatives, in cases of hepatic encephalopathy, gave written informed consent to participate in the study. It was approved by the Clinical Investigation and Ethics Committee of the Hospital Clinic of Barcelona. On resolution of hepatic encephalopathy, informed consent was requested from the patients for continuation in the study.
Inclusion and the baseline clinical evaluation was performed within 24 hours of hospitalization and included history and physical examination, liver and renal tests, ascitic fluid analysis and culture, fresh urine sediment, chest x-ray, and abdominal ultrasonography. Heart and respiratory rates and body temperature were recorded to estimate systemic inflammatory response syndrome (SIRS). Mean arterial pressure, calculated as the median of three values, was measured noninvasively with the patient in supine position with a 5-minute interval (DINAMAP Vital Signs Monitor, Critikon, Tampa, FL). Severity of liver failure was estimated by the Child-Pugh and the model for endstage liver disease (MELD) scores. Fasting blood samples were also obtained within this first 24 hours after hospital admission for assessment of vasoactive mediators, proinflammatory cytokines, and lipid profile. Samples were obtained in all patients through an intravenous catheter inserted at least 6 hours before sampling.
Adrenal Function Tests
A short Synacthen test (SST) was performed between 8:00 and 9:00 am within the first 24 hours of admission. Synthetic adrenocorticotropic hormone (250 μg, Synacthen, Novartis Pharma, Basel, Switzerland) was given intravenously. Blood samples to measure serum total cortisol levels (competitive immunoassay using direct chemiluminescent technology; Advia-Centaur, Bayer, Pittsburgh, PA) were obtained prior and 60 minutes following Synacthen administration. The coefficient of variation for this test is 7%. Baseline serum levels of transcortin were also determined (competitive radioimmunoassay; Biosource, Belgium; interassay coefficient of variation: 5.5%). Results of the SST were available to the treating physicians.
Free cortisol was not directly tested. We estimated its concentration using the free cortisol index (FCI) and the calculated free cortisol (cFC). FCI is the ratio between total cortisol (nmol/L) and transcortin (mg/L). cFC was derived using Coolens' equation: U2K(1 + N) + U[1 + N + K(T − C)] − C = 0, where U represents the molar concentration of unbound cortisol, C is the molar concentration of total cortisol, T is the concentration of trancortin, and K is the affinity of transcortin for cortisol at 37°C. N is the ratio of albumin bound to free cortisol, and 1.74 is the value conventionally used.
Activity of Vasoactive Systems, Cytokines, and Lipid Profile
Baseline plasma renin activity (patient in supine position for at least 1 hour), plasma concentrations of vasopressin, norepinephrine, interleukin-6 (IL-6), tumor necrosis factor alpha (TNF-α), and serum levels of nitric oxide, total serum cholesterol and triglyceride, and high density lipoprotein (HDL) and low density lipoprotein (LDL) cholesterol were determined at inclusion. Plasma renin activity and plasma concentration of vasopressin and norepinephrine were assessed by radioimmunoassay (Clinical Assays, Cambridge, MA; Bühlman Laboratories, Basel, Switzerland, and CAIBL Laboratories, Hamburg, Germany, respectively). To measure serum levels of nitrates and nitrites (NO2− and NO3−), samples were ultrafiltered (PL-10 Ultrafree-MC centrifugal filter units; Millipore, Bedford, MA) at 1,200g for 1 hour to remove proteins before analysis. Filtered serum was refluxed in glacial acetic acid containing sodium iodide. Under these conditions NO2− and NO3− are reduced to NO, which, after reacting with ozone, can be quantified by a chemiluminescence detector (Nitric Oxide Analyzer, NOA 280, Sievers Instruments, Boulder, CO). IL-6 and TNF-α were measured by enzyme-linked immunosorbent assay (Medgenix Diagnostics, Fleurus, Belgium). Serum levels of total and HDL cholesterol and triglyceride were measured by the enzymatic colorimetric methods (ADVIA 2400 Chemistry System; Bayer Health Care). LDL cholesterol was calculated using the Friedewald formula.
Normal values in our laboratory are: transcortin 25-55 μg/mL, plasma renin activity: 1.4 ± 0.9 ng/mL*h, norepinephrine: 136-364 pg/mL, vasopressin: 1.5-3.3 ng/L, IL-6: < 5 pg/mL, TNF-alpha: < 20 pg/mL, nitrates and nitrites: 37 ± 14 nMol/mL, total cholesterol: 148-247 mg/dL, triglycerides: 50-150 mg/dL, HDL cholesterol: >40 mg/dL, and LDL cholesterol: <180 mg/dL. FCI above 12 represents sufficient adrenal function.
All patients were managed following standard protocols for each clinical decompensation. Patients were followed during hospitalization and monthly up to 3 months. Any significant new clinically relevant event was recorded including bacterial infections, gastrointestinal bleeding, hepatic encephalopathy, and HRS. New clinical events developed during hospitalization and up to 3 months, but not those diagnosed at inclusion, were included in the analysis of clinical outcome and mortality of patients with and without RAI. Severity of infections (severe sepsis and septic shock) and of gastrointestinal bleeding (presence of shock) occurring during follow-up was also evaluated. Finally, hospital and 3-month mortality and causes of death were recorded.
This was an observational study and the protocol did not consider the administration of hydrocortisone in patients with RAI. Only patients developing septic shock during follow-up (three with RAI and one without RAI at inclusion) received stress dose steroids according to our current guidelines. The rest of the patients did not receive steroids.
RAI was diagnosed when the increase in serum total cortisol after SST was <9 μg/dL in patients with basal serum total cortisol <35 μg/dL. We chose this diagnostic criteria for two reasons: (1) it is the gold standard criteria used in critical care, the setting where this entity was first described, and (2) because it is not affected by changes in transcortin or albumin levels, thus avoiding overdiagnosis of RAI due to falsely low serum total cortisol levels in patients with advanced cirrhosis[10, 11] Diagnosis of different bacterial infections was done according to criteria previously reported.[25, 26] Patients were considered to have SIRS (sepsis) if they fulfilled at least two of the following criteria: (i) a core temperature >38°C or <36°C; (ii) a heart rate >90 beats/min; (iii) a respiratory rate >20 breaths/min in the absence of hepatic encephalopathy; and (iv) a white blood cell count >12,000 or <4,000 /mm3, or a differential count showing ≥10% immature polymorphonuclear neutrophil cells. Severe sepsis was defined by the presence of SIRS and an acute organ failure. Septic shock was diagnosed by the presence of data compatible with SIRS, mean arterial pressure below 60 mmHg for more than 1 hour despite adequate fluid resuscitation (increase in central venous pressure to 8-10 mmHg), and need of vasopressor drugs. Hypovolemic shock was defined by the presence of bleeding and a systolic pressure <90 mmHg and heart rate >100 b/min. Type-1 HRS was diagnosed according to the International Ascites Club (IAC) criteria. Acute-on-chronic liver failure (ACLF) was defined according to the criteria recently established in the CANONIC study. Consequently, this specific cause of death was retrospectively identified.
Differences were considered significant at 0.05. Results are given as mean (SD). Continuous variables were compared by the Student t test or by the Mann-Whitney U test when indicated. Discontinuous variables were compared by the chi-squared test. Yates' correction was applied when the number of cases in a cell was lower than five. Probability curves were obtained by the Kaplan-Meier method and compared by the log-rank test. Multivariate analyses were made using logistic regression. Calculations were performed with the SPSS Statistical Package (v. 18.0, Chicago, IL).
Study Population and Clinical Characteristics of Patients
A total of 658 cirrhosis patients with acute decompensation requiring hospitalization were screened during the 20-month study period. Of these, 515 were not included due to the presence of exclusion criteria (422), death between evaluation and baseline analysis (16), or refusal to participate (77) (Supporting Fig. 1). The study was therefore performed in 143 patients. The main cause of admission in the series was infection in 61 patients (43%), followed by variceal bleeding in 29 (20%), ascites in 27 (19%), hepatic encephalopathy in 11 (8%), HRS in 8 (6%), and other causes in 7 (5%). The most common infection at inclusion was spontaneous bacterial peritonitis (26), followed by cellulitis (10), urinary tract infection (8), and pneumonia (6).
Seventy-two percent of patients were men. The mean age was 57 ± 9 years. The cause of cirrhosis was alcoholism in 73 cases, hepatitis C virus (HCV) in 40, HCV plus alcohol in 20, hepatitis B virus in five, and other causes in five. Most patients were severely ill as indicated by the poor hepatic and renal function. The mean Child-Pugh and MELD scores were 9.39 ± 2.14 and 18.21 ± 6.75, respectively. In all, 102 patients had ascites, 44 hepatic encephalopathy, and 34 gastrointestinal hemorrhage. Eight patients were in the intermediate critical care area at inclusion, seven due to variceal bleeding and one because of grade 3 hepatic encephalopathy.
Clinical characteristics at admission of patients included in the study were similar to those of patients who refused to participate or were excluded because of >24 hours from admission (data not shown).
Prevalence, Clinical and Laboratory Data of Patients With and Without RAI
RAI was diagnosed in 37 patients of the series (26%). Prevalence of adrenal dysfunction did not significantly differ regarding the presence or absence of specific clinical decompensations at inclusion: ascites (28% versus 20%, respectively), hepatic encephalopathy (30% versus 24%), variceal bleeding (19% versus 28%), bacterial infection (19% versus 32%), and SBP versus non-SBP infections (15% versus 22%). Only patients with type-1 HRS showed a trend towards a higher prevalence of RAI (57% versus 24%, P = 0.07). The prevalence of RAI was also similar across different Child-Pugh classes: 21% in Child-Pugh class A, 25% in class B and 28% in class C patients (P = 0.87).
Table 1 shows the clinical and analytical characteristics of patients with and without RAI at inclusion in the study. Patients with RAI presented poorer renal function (higher blood urea nitrogen [BUN] levels and lower serum sodium concentration) and higher degree of circulatory dysfunction (lower mean arterial pressure) than patients with normal adrenal function. Liver function (Child-Pugh and MELD scores), type of decompensations (ascites, hepatic encephalopathy, hemorrhage, or bacterial infection), and inflammatory markers (serum C reactive protein levels and blood leukocyte count) did not differ between patients with normal and abnormal adrenal function.
|RAI (n = 37)||Normal Adrenal Function (n = 106)||P|
|Age (years)||55 ± 9||57 ± 10||NS|
|Male sex (%)||68||74||NS|
|HCV cirrhosis (%)||48||53||NS|
|Active alcoholisma (%)||32||32||NS|
|Long-term antibiotic prophylaxis (%)||24||22||NS|
|Bacterial infection (%)b||35||51||NS|
|Hepatic encephalopathy (%)||35||29||NS|
|Gastrointestinal bleeding (%)||19||26||NS|
|Type 1 HRS (%)||11||3||0.07|
|Mean arterial pressure (mmHg)||76 ± 12||83 ± 14||0.009|
|Heart rate (beats/min)||86 ± 12||86 ± 17||NS|
|Leukocyte blood count (x109/L)||7.5 ± 5.4||7.3 ± 5.2||NS|
|SIRS criteria (%)c||60||41||0.08|
|Serum bilirubin (mg/dL)||6.1 ± 8.4||4.4 ± 4.6||NS|
|Prothrombin time (%)||49 ± 15||50 ± 14||NS|
|Serum creatinine (mg/dL)||1.33 ± 1.06||1.18 ± 1.04||NS|
|BUN (mg/dL)||32 ± 24||24 ± 15||0.06|
|Serum sodium (mEq/L)||131 ± 7||135 ± 5||0.007|
|C reactive protein (mg/dL)||4.5 ± 5.2||3.6 ± 4.2||NS|
|Child-Pugh score (points)||9.7 ± 2.1||9.3 ± 2.2||NS|
|MELD score (points)||19.4 ± 7.9||17.8 ± 6.2||NS|
|Intermediate care unit (%)e||8||5||NS|
Vasoactive Systems and Cytokine Profile
Baseline plasma renin activity and plasma concentration of norepinephrine were significantly higher in patients with RAI compared to those with normal adrenal function, indicating a higher degree of circulatory dysfunction and homeostatic activation of the renin-angiotensin and sympathetic nervous systems (Table 2). Plasma cytokines levels (IL-6 and TNF-α) were higher in patients with RAI, although the difference was not statistically significant due to the high variation in cytokine levels. No significant differences were observed between groups regarding plasma levels of vasopressin and serum levels of nitric oxide.
|RAI (n = 37)||Normal Adrenal Function (n = 106)||P|
|Plasma renin activity (ng/mL*h)||7.1 ± 9.9||3.4 ± 5.6||0.03|
|Plasma norepinephrine (pg/mL)||544 ± 334||402 ± 316||0.02|
|Plasma vasopressin (ng/L)||3.0 ± 2.0||3.2 ± 3.7||NS|
|Plasma TNF-α (pg/mL)||54 ± 115||27 ± 24||NS|
|Plasma IL-6 (pg/mL)||916 ± 2532||244 ± 439||NS|
|Serum nitric oxide (nmol/mL)||59 ± 65||47 ± 62||NS|
Adrenal Axis and Cholesterol Profile
Table 3 shows serum total cortisol levels before and after the SST, transcortin, and albumin levels and serum cholesterol profile in patients with and without RAI. By definition delta cortisol and post-SST cortisol levels were significantly lower in patients with RAI. Baseline serum total cortisol levels, serum levels of transcortin (the main cortisol binding protein), albumin, total cholesterol, and HDL were not significantly different between patients with normal and abnormal adrenal function. LDL levels tended to be lower in patients with RAI. Estimated baseline free cortisol levels (FCI and cFC) were also similar between groups.
|RAI (n = 37)||Normal Adrenal Function (n = 106)||P|
|Serum total cortisol levels|
|Basal cortisol (μg/dL)||16.8 ± 5.5||15.3 ± 6.4||NS|
|Post ACTH cortisol (μg/dL)||23.0 ± 5.5||29.9 ± 7.4||0.0001|
|Delta cortisol (μg/dL)||6.1 ± 2.6||14.6 ± 4.4||0.0001|
|Cortisol binding protein levels|
|Transcortin (μg/dL)||30.2 ± 10.5||28.4 ± 9.4||NS|
|Serum albumin (g/L)||27.9 ± 4.5||28.5 ± 5.3||NS|
|Estimated free cortisol|
|Free cortisol index (nmol/mg)||16.5 ± 6.4||16.4 ± 8.6||NS|
|Coolens formula||0.18 ± 0.43||0.29 ± 0.91||NS|
|Total cholesterol (mg/dL)||93 ± 36||106 ± 43||NS|
|HDL (mg/dL)||21 ± 16||23 ± 12||NS|
|LDL (mg/dL)||57 ± 25||67 ± 33||0.08|
Reevaluation of Adrenal Axis
In 18 patients (3 with and 15 without RAI) SST was repeated 153 ± 151 days after inclusion. Two out of the three patients with RAI and 14 out of the 15 patients with normal adrenal function at admission showed normal delta values at follow-up. These data suggest that adrenal function in cirrhosis patients without RAI is relatively stable and that RAI is potentially reversible.
Clinical Outcome and Mortality of Patients With and Without RAI
Mean duration of hospitalization was 13 ± 12 days (from 2 to 83 days) with no significant differences between patients with and without RAI. Clinical outcome differed significantly between patients with normal and abnormal adrenal function (Table 4). The probability of developing new bacterial infections (24% versus 9%; P = 0.01), new episodes of severe sepsis or septic shock (19% versus 4%, P = 0.008), and new type-1 HRS (11% versus 1%, P = 0.006) was significantly higher in patients with RAI than in those with normal adrenal function. The probability of death during hospitalization (16% versus 4%, P = 0.02) was also higher in patients with RAI. No new episodes of variceal bleeding occurred during hospitalization in either group.
|New Clinical Eventa||RAI (n = 37)||Normal Adrenal Function (n = 106)||Pb||Log Rank|
|During hospitalization (n, %)|
|Bacterial infection||9 (24%)||10 (9%)||0.045||0.01|
|SBP||3 (8%)||2 (2%)||NS||NS|
|Non-SBP infection||6 (16%)||8 (8%)||NS||0.03|
|Severe sepsis or septic shock||7 (19%)||4 (4%)||0.007||0.005|
|Septic shock||2 (5%)||—||0.07||0.04|
|Type-1 HRSc||4 (11%)||1 (1%)||0.02||0.006|
|Death||6 (16%)||4 (4%)||0.02||0.02|
|At 3 months (n, %)|
|Variceal bleeding||—||5 (5%)||NS||NS|
|Bacterial infection||15 (41%)||22 (21%)||0.03||0.008|
|SBP||4 (11%)||5 (5%)||NS||NS|
|Non SBP infection||11 (30%)||17 (16%)||NS||0.02|
|Severe sepsis or septic shock||10 (27%)||9 (9%)||0.009||0.003|
|Septic shock||3 (8%)||3 (3%)||NS||NS|
|Type-1 HRSc||6 (16%)||3 (3%)||0.01||0.002|
|Death||8 (22%)||7 (7%)||0.02||0.01|
Evolution at 3 Months
Mean follow-up was similar in patients with and without RAI (72 ± 30 versus 78 ± 25 days, respectively). Main outcomes at 3 months also differed between patients with normal and abnormal adrenal function (Table 4). The 3-month probability of developing new bacterial infections (41% versus 21%; P = 0.008), new severe sepsis, or septic shock episodes (27% versus 9%, P = 0.003, Fig. 1) and new type-1 HRS (16% versus 3%, P = 0.002) was higher in patients with RAI than in those with normal adrenal function. The probability of death was also significantly higher in patients with RAI (22% versus 7%, P = 0.01, Fig. 2). New episodes of variceal bleeding (n = 5) occurred only in patients without RAI.
Causes of Death
ACLF was the main cause of death in patients with and without RAI both during hospitalization (five versus two, respectively) and at 3 months (six versus four, respectively). Septic shock (two and one, respectively) and respiratory failure (two patients with normal adrenal function) were responsible for the remaining deaths at 3 months.
Risk Factors for the Development of Severe Sepsis, Type-1 HRS, and for Death at Short-Term
Table 5 shows factors associated to the development of severe sepsis, type-1 HRS, and death at 3 months in the univariate analysis. Considering the low number of events observed in the study we decided to include only four of the variables with significant predictive value in each of the multivariate models: MELD score, which reflects hepatic and renal function, both plasma renin activity and plasma noradrenaline concentration as markers of circulatory dysfunction, and delta cortisol as an estimation of adrenal function. Table 6 shows the independent predictors in the different models. Delta cortisol, a dynamic marker of adrenal function, was identified as independent risk factor of all three short-term outcomes (severe sepsis, type-1 HRS, and mortality).
|Severe sepsis or septic shock||Yes||No|
|n = 19||n = 124||P|
|BUN (mg/dL)||40 ± 25||24 ± 15||0.002|
|Serum sodium (mEq/L)||130 ± 6||134 ± 5||0.002|
|MELD score (points)||23 ± 10||18 ± 6||0.02|
|Delta cortisol (μg/dL)||8.9 ± 3.8||13.0 ± 5.5||0.002|
|Plasma renin activity (ng/mL*h)||10.6 ± 11.2||3.4 ± 5.8||0.0001|
|Plasma noradrenaline (pg/mL)||605 ± 315||415 ± 321||0.02|
|Type-1 HRS||n = 9||n = 134||P|
|Ascites at inclusion (%)||100||69||0.06|
|Serum sodium (mEq/L)||130 ± 7||134 ± 5||0.06|
|Child-Pugh score (points)||11 ± 2||9 ± 2||0.01|
|MELD score (points)||25 ± 8||18 ± 6||0.003|
|Delta cortisol (μg/dL)||8.7 ± 3.2||12.7 ± 5.5||0.04|
|Total serum cholesterol (mg/dL)||77 ± 22||105 ± 42||0.02|
|Plasma renin activity (ng/mL*h)||8.6 ± 7.7||4.1 ± 7.0||0.01|
|Death||n = 15||n = 128||P|
|Hepatic encephalopathy (%)||67||27||0.003|
|BUN (mg/dL)||37 ± 21||25 ± 17||0.02|
|Serum sodium (mEq/L)||130 ± 7||134 ± 5||0.02|
|Serum albumin (g/L)||25 ± 4||29 ± 5||0.006|
|Child-Pugh score (points)||11 ± 2||9 ± 2||0.0001|
|MELD score (points)||26 ± 9||17 ± 6||0.0001|
|Delta cortisol (μg/dL)||8.1 ± 3.9||12.9 ± 5.4||0.001|
|Total serum cholesterol (mg/dL)||77 ± 27||106 ± 42||0.01|
|HDL cholesterol (mg/dL)||11 ± 12||24 ± 13||0.001|
|Plasma noradrenaline (pg/mL)||833 ± 474||396 ± 275||0.004|
|Severe sepsis or septic shock|
|Plasma renin activity||1.07||1.03-1.11||0.001|
Our results indicate that nearly one-fourth of noncritically ill patients with cirrhosis admitted to the hospital for the treatment of acute decompensation present RAI. Among the different methods currently available to assess adrenal function (measurements of baseline total or free cortisol levels in serum, plasma, or saliva and changes in cortisol after insulin-induced hypoglycemia or the administration of 1 or 250 μg of adrenocorticotropic hormone [ACTH] or 1 μg/kg of corticotropin-releasing hormone)[22, 32-35] we chose the SST (increase in total serum cortisol levels 1 hour after the administration of 250 μg of ACTH, Synacthen) because it is the gold standard test used to define this entity in critical care, the setting where RAI was first described. It is also a dynamic test routinely used in the evaluation of adrenal function in clinical endocrinology.[36, 37] Among the possible criteria that can be used to define RAI using the SST: baseline serum total cortisol levels, peak serum total cortisol levels, delta cortisol, or a combination of them, we decided to use only the delta value, because as dynamic criteria it is not affected by changes in transcortin or albumin levels. Furthermore, several studies have shown that serum total cortisol overestimates the prevalence of RAI in cirrhosis due to low transcortin and albumin concentrations.[17-20] Although free cortisol levels might estimate more adequately the real prevalence of RAI in the cirrhosis population, they are not routinely used because the determination technique is complex and expensive and because diagnostic cutoff values have not been clearly defined.
The mechanism of RAI in cirrhosis is probably multifactorial. It has been suggested that it could be the result of increasing circulating levels of endotoxin and other bacterial products, which increase proinflammatory cytokine concentration and depress the adrenal synthesis of steroids.[10, 22] Liver failure, by decreasing the production of cholesterol, the cortisol substrate, could also play a contributing role. Finally, adrenal hypoperfusion secondary to the circulatory dysfunction of cirrhosis could be also involved in the pathogenesis of RAI.[10, 22] In the current study we found that patients with RAI had higher plasma renin, noradrenaline, and cytokine levels and lower serum levels of LDL cholesterol than patients without RAI, although significant differences were only observed in relationship to the endogenous vasoactive hormones.
An interesting finding of the study was that the prevalence of RAI did not correlate with the severity of liver disease. This finding suggests that factors other than liver function (adrenal perfusion, inflammatory response, nutritional parameters) can be involved in its pathogenesis.
Serum cortisol is an essential component in the homeostasis of circulatory function in humans since it increases the vascular and cardiac responses to angiotensin-II and catecholamines.[1, 2, 38] Our study shows that RAI is associated with a significantly lower mean arterial pressure and higher plasma renin activity and norepinephrine concentration. RAI was also associated with a more deteriorated glomerular filtration rate and renal free water excretion, as indicated by higher BUN and lower serum sodium levels. Taken together, these results suggest that RAI is associated with impairment in circulatory and renal function.
A relevant finding of our study was the association between RAI and the risk to develop new infections, severe sepsis, and type 1 HRS. It is well known that circulatory dysfunction and the secondary activation of the sympathetic nervous system impairs several defensive mechanisms against enteric infections.[39, 40] Overactivation of the sympathetic nervous system inhibits intestinal motility and increases bacterial overgrowth. The increased release of catecholamines from adrenergic terminals exert potent immunosuppressive actions including inhibition of chemotaxis/migration and phagocytosis of bacteria by neutrophils and monocytes.[42, 43] Finally, catecholamines are released into the intestinal lumen, where they interact with specific bacterial receptors and increase bacterial growth, adherence to the mucosa, penetration into the interstitial space and lymphatics within the intestinal wall, and virulence.[40, 44] The net effect of all these alterations is an increased translocation of bacteria and bacterial products from the intestinal lumen to the submucosal lymphatics and then to the systemic circulation, giving rise to spontaneous bacterial infections and systemic inflammatory response. Since adrenal dysfunction is associated with reduction in the vascular effect of the renin-angiotensin and the sympathetic nervous systems, the compensatory increase in the adrenergic tone could contribute to increase the risk of bacterial infection.
Circulatory and renal dysfunction and overactivity of the renin-angiotensin and sympathetic nervous systems are well-known risk factors of HRS development in patients with decompensated cirrhosis. On the other hand, bacterial infections in cirrhosis are frequently associated to the development of type-1 HRS.[45, 46] These factors could account for the higher risk of type-1 HRS observed in patients with RAI. Another interesting observation of our study was that patients with RAI and bacterial infection developed more frequently severe sepsis or shock. The more severe impairment in circulatory function prior to infection and perhaps an exaggerated inflammatory response due to low circulating cortisol levels could account for this feature.
The probability of death was significantly higher in noncritically ill patients with RAI than in those with normal adrenal function. The main cause of death was ACLF, a recently defined syndrome in patients with acute decompensation of cirrhosis characterized by one or more organ failures, intense systemic inflammatory response, and very high mortality. The second cause of mortality in this series was septic shock.
In the analysis of independent risk factors for the development of severe sepsis, type-1 HRS, and death, delta cortisol together with three important predictive variables (MELD, which estimates the degree of liver and renal dysfunction, and plasma renin activity and norepinephrine concentration, which estimate the degree of circulatory dysfunction) were introduced in the models. Delta cortisol and MELD were found to be independent predictors of severe sepsis, type-1 HRS, and mortality. Plasma renin activity and plasma noradrenaline were also independent risk factors of severe sepsis and death, respectively.
A potential weakness of our study is the heterogeneity of patients included. The study was designed to evaluate the prevalence of RAI and its relationship to clinical course in noncritically ill cirrhosis patients with acute clinical decompensation, thereby including subjects with ascites, encephalopathy, bacterial infection, variceal bleeding, or HRS. Although the prevalence of RAI did not significantly differ among different patient groups (except for type-1 HRS), mechanisms of adrenal dysfunction and its association with clinical events may differ among different decompensations of cirrhosis. Further studies should clarify this point.
In summary, our study shows that RAI is a relatively frequent event in noncritically ill cirrhosis patients with acute decompensation and appears to be associated with impairment in circulatory and renal function and higher risk of short-term development of bacterial infections, severe sepsis, type-1 HRS, and death.
- 7American College of Critical care Medicine. Recommendations for the diagnosis and management of corticosteroid insufficiency in critical 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., , , , , , et al;
- 35Comparison of the low dose short synacthen test (1 microg), the conventional dose short synacthen test (250 microg) and the insulin tolerance test for assessment of the hypothalamo-pituitary-adrenal axis in patients with pituitary disease. J Clin Endocrinol Metab 1999;84:838-843., , , .
- 37Evaluation of adrenal function in patients with growth hormone deficiency and hypothalamic-pituitary disorders: comparison between insulin-induced hypoglycemia, low-dose ACTH, standard ACTH and CRH stimulation tests. Eur J Endocrinol 2005;152:735-741., , , , , , et al.
- 40Role of sympathetic nervous activity for bacterial translocation in advanced experimental liver cirrhosis. Gastroenterol Hepatol 2011;34:102-105..
Additional Supporting Information may be found in the online version of this article.
|hep26535-sup-0001-suppfig1.tif||10172K||Supporting Figure 1 Legend Flow diagram of patients included in the study. HIV: human immunodeficiency virus; COPD: chronic obstructive pulmonary disease.|
|hep26535-sup-0002-suppfig2.tif||10172K||Supporting Information Figure 2|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.