Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival

Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival

G Model YDLD-2877; No. of Pages 7 ARTICLE IN PRESS Digestive and Liver Disease xxx (2015) xxx–xxx Contents lists available at ScienceDirect Digesti...

667KB Sizes 0 Downloads 46 Views

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS Digestive and Liver Disease xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Digestive and Liver Disease journal homepage: www.elsevier.com/locate/dld

Liver, Pancreas and Biliary Tract

Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival Alessandro Risso a,∗ , Carlo Alessandria a , Lavinia Mezzabotta a , Chiara Elia a , Alida Andrealli a , Maurizio Spandre a , Paola Di Luigi a , Anna Barbui b , Andrea Evangelista c , Anna Morgando a , Roberto Serra b , Giovannino Ciccone c , Alfredo Marzano a , Mario Rizzetto a a

Division of Gastroenterology and Hepatology, AOU Città della Salute e della Scienza di Torino, University of Turin, Italy Laboratory of Microbiology, Department of Clinical Pathology, AOU Città della Salute e della Scienza di Torino, University of Turin, Italy c Unit of Cancer Epidemiology, CPO Piemonte, AOU Città della Salute e della Scienza di Torino, University of Turin, Italy b

a r t i c l e

i n f o

Article history: Received 11 November 2014 Accepted 16 April 2015 Available online xxx Keywords: Adrenal insufficiency Ascites Bacterial DNA Cirrhosis

a b s t r a c t Background: There are few data on clinical relevance of adrenal dysfunction and its relationship with occult microbial DNA in noninfected haemodynamically stable cirrhotic patients with ascites. Aims: The aim of this study was to evaluate prognostic role of adrenal dysfunction, microbial DNA, and their relationship. Methods: Adrenal function was assessed in 93 consecutive patients following a corticotropin stimulation test. Adrenal dysfunction was defined as: basal cortisol <10 ␮g/dl, delta cortisol <9 ␮g/dl, or peak cortisol <18 ␮g/dl. Microbial DNA was assessed in blood and ascites of 54 consecutive patients. Patients were followed up until liver transplantation or death. Results: Adrenal dysfunction was not significantly associated with mortality, while the risk of death rose significantly with an increase in basal cortisol values (HR 1.13 per 1-␮l/dl increase; 95% CI 1.01–1.26). Microbial DNA was independently associated with reduced survival (HR 8.05, 95% CI 1.57–41.2). In microbial DNA-positive patients a significant correlation was found between Model for End-Stage Liver Disease (MELD) score and basal cortisol values (Pearson’s r = 0.5107; p = 0.018). Conclusions: Microbial DNA and MELD score, but not adrenal function, were the best independent predictors of mortality in noninfected cirrhotic patients with ascites. High serum cortisol levels may be a systemic reaction to microbial translocation, increasing in parallel with deterioration of liver function. © 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

1. Introduction Adrenal dysfunction (AD) is defined as inadequate cellular corticosteroid activity for the severity of the patient’s illness [1]. It is common in patients with septic shock, and it is associated with circulatory collapse and hyporesponsiveness to vasopressors [2,3]. Intravenous administration of hydrocortisone reduces the time of shock reversal and the need for vasopressors; however, data on mortality are contradictory [4–6].

∗ Corresponding author at: Division of Gastroenterology and Hepatology, AOU Città della Salute e della Scienza di Torino, University of Turin, Corso Bramante 88, Turin 10126, Italy. Tel.: +39 011 6335561/5569/6397; fax: +39 011 6335714/5927. E-mail address: [email protected] (A. Risso).

AD has also been investigated in patients with liver disease, adopting the same definitions used in the general population of critically ill patients [1–3]. Most studies were performed in patients with acute liver failure or septic shock: in this context, AD is common and it is associated with higher mortality rates. Hydrocortisone infusion showed encouraging, although non-univocal, results [7–11]. In patients with advanced cirrhosis without sepsis or other acute distress, data are limited [12–16]. A very recent study confirmed the high prevalence and the relevance of AD in noncritically ill cirrhotic patients, as patients with AD are at higher risk of circulatory and renal dysfunction, sepsis, and death [12]. In cirrhosis, AD was related to deficient hepatic synthesis of cortisol precursors (high-density lipoprotein (HDL) cholesterol and A1 apolipoprotein) due to liver insufficiency together with high

http://dx.doi.org/10.1016/j.dld.2015.04.010 1590-8658/© 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

2

levels of circulating inflammatory cytokines, secondary to bacterial translocation from the gut [1,17]. This relationship was further investigated in recent years with the introduction of bacterial DNA measurement in blood and ascites to assess, at a molecular level, the presence of microorganisms in apparently noninfected cirrhotic patients [18–22], and with the demonstration that bacterial DNA is an independent predictor of poor survival in these patients [19]. Similar findings were also suggested for fungal DNA [23]. This prospective study evaluated the prognostic value of adrenal function, the presence and the clinical relevance of microbial DNA, and their relationship in noninfected cirrhotic patients with ascites. 2. Methods All consecutive patients with cirrhosis and ascites admitted to the Gastroenterology and Hepatology Unit (19 beds) of our academic hospital between August 2008 and April 2011 were considered for inclusion. The diagnosis of cirrhosis was based on clinical, laboratory, and ultrasonographic findings. The inclusion criteria were as follows: cirrhosis with noninfected ascites (polymorphonuclear leukocytes <250/␮l and negative ascitic fluid cultures), age between 18 and 75 years, and ability to provide written informed consent. The exclusion criteria were: shock, infection, or signs of bleeding within a week prior to inclusion; treatment with steroids or vasoactive drugs; multifocal hepatocellular carcinoma; alcoholic hepatitis; intrinsic renal failure; and heart or respiratory failure. Patients underwent a clinical workup; blood, urine, and ascites were collected from all patients. Hepatic and renal function, activation of neurohormonal endogenous systems, and the levels of steroid synthesis precursors were determined. The presence of an underlying infection was ruled out by white blood cell count; paracentesis with polymorphonuclear leucocyte count; ascites, blood, and urine cultures; chest X-ray; and ultrasonography. Patients were followed up until liver transplantation (LT) or death. All procedures were in accordance with the ethical standards of committees on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008. Written informed consent was obtained from all patients. 2.1. Assessment of adrenal function Patients underwent a short corticotropin stimulation test (SST) [24,25]. The serum total cortisol level was assessed at 8.00 a.m.; then, 250 ␮g of synthetic adrenocorticotrophic hormone (ACTH) (tetracosactide: Synachten, NOVARTIS FARMA S.p.A., Naples, Italy) was intravenously administered, and the serum total cortisol level was redetermined 60 min after the infusion. 2.2. Definitions of AD Adrenal function was determined from the values of serum total cortisol before and 60 min after the administration of synthetic ACTH, using the following three standard criteria of AD: (a) “Basal cortisol”: serum total cortisol before SST <10 ␮g/dl [1,2]; (b) “Delta cortisol”: difference between serum total cortisol after and before SST <9 ␮g/dl [7,9]; (c) “Peak cortisol”: serum total cortisol after SST <18 ␮g/dl [8,10].

(Roche Diagnostics, Meylan, France). The intra-assay coefficient of variation (CV) is 1.4% in the cortisol range of 4–20 ␮g/dl. The plasma renin activity (PRA) and plasma aldosterone levels were measured by radioimmunoassay. The normal values for PRA and plasma aldosterone in our laboratory are 0.1–4 ng/ml/h and 12–150 pg/ml, respectively. The C-reactive protein (CRP) normal range in our laboratory is 0–3 mg/l. 2.4. Bacterial and fungal DNA Blood and ascitic fluid were collected from all patients for bacterial and fungal cultures by inoculation of 10 ml of the sample in aerobe and anaerobe culture medium of the BactAlert automatic system (Biomerieux, Lyon, France) for 5 days. Determination of bacterial and fungal DNA in serum, blood, and ascites was also performed as soon as microbial DNA assay became available in our hospital (October 2009). Microbial DNA was extracted from 200 ␮l of serum or from the pellet of 5 ml of ascitic fluid using the EZ1 automatic system (Qiagen, Hilden, Germany), and from 1 ml of whole blood with the silica column manual system (Molzym, Bremen, Germany). The serum and ascitic fluid samples were eluted in 50 ␮l of elution buffer, while the whole blood samples were eluted in 100 ␮l of elution buffer, and 10 ␮l of the extracted DNA was examined for bacterial and fungal DNA by polymerase chain reaction (PCR). Bacterial DNA was detected by amplifying a 500-bp region at the 5 end of 16 S ribosomal RNA (rRNA) gene using a MicroSeq500 16sDNA Bacterial Identification PCR kit, and the sequencing reactions were performed with the MicroSeq500 16sDNA Bacterial Identification Sequencing kit (Applied Biosystems). In brief, 10 ␮l of the extracted DNA was mixed with the reaction mixture provided in the kit and amplified. The cycling conditions used were as follows: 40 cycles of 95 ◦ C for 30 s, 60 ◦ C for 30 s, and 72 ◦ C for 45 s. To determine the presence of a PCR product, 5 ␮l of the PCR reaction was run in a 2.0% agarose gel. The size of the PCR product was about 500 bp (460–560 bp, depending on the bacterial species). Fungal DNA was detected by amplifying the genomic region internal transcribed spacer 2 (ITS2) rDNA (ITS1: 5 -TCC GTA GGT GAA CCT GCG G-3 and ITS4: 5 -TCC TCC GCT TAT TGA TAT GC-3 ). In brief, 10 ␮l of the extracted DNA was mixed with a reaction PCR mixture containing the enzyme and buffers (Master Mix PCR Qiagen) and 25 pmol of each primer; the cycling conditions used were as follows: 40 cycles of 95 ◦ C for 60 s, 42 ◦ C for 60 s, and 72 ◦ C for 60 s. To determine the presence of a PCR product, 5 ␮l of the PCR reaction was run in a 2.0% agarose gel. The size of the PCR product was about 700 bp, depending on the species. The absence of PCR inhibitors was confirmed by ␤-globin gene PCR detection in all extracted DNA samples. DNA extraction and detection of bacterial and fungal DNA were performed using 10 samples of serum and 10 samples of whole blood from blood donors to control the sterility of sampling, DNA extraction, and DNA amplification protocols. The PCR products were identified by sequencing the amplicons obtained. The nucleotide sequences of the PCR products were determined using an ABI PRISM Dye Terminator Cycle Sequencing v2.0 Ready Reaction Kit (Perkin Elmer, Foster City, CA, USA) and an ABI PRIMS 377 automated sequencer. The sequences were identified using BLAST at the National Center for Biotechnology Information web site (www.ncbi.nlm.nhi.gov).

2.3. Methods of measurement 2.5. Statistical analysis The normal values of serum total cortisol in our laboratory are 4–20 ␮g/dl. The serum total cortisol level was measured using electrochemiluminescence-based methods on the Cobas e601

The cumulative incidence of death (CID) was determined from the date of SST to the date of death or of LT. Patients who

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

3

Table 1 Demographic, clinical and laboratory data at inclusion for the total cohort of patients and for patients undergoing microbial DNA assessment. Total cohort (93 patients)

Patients assessed for microbial DNA (54 patients) Negative microbial DNA (33 patients)

Positive microbial DNA (21 patients)

p

Age (years) Alcohol abuse (n, %)

57 (50–66) 39 (42%)

57 (53–66) 15 (45%)

61 (47–65) 9 (43%)

0.85 0.85

Aetiology of cirrhosis (n, %) Alcohol abuse only HBV HCV HBV + HCV HBV + HDV HBV + HCV + HDV Autoimmune hepatitis Primary biliary cirrhosis Primary sclerosing cholangitis Haemochromatosis Cryptogenic cirrhosis Previous hepatic encephalopathy (n, %) Previous upper gastrointestinal bleeding (n, %) Previous SBP (n, %) Previous severe infection/sepsis (n, %) Previous type-1 HRS (n, %) Child–Pugh score: points MELD score (points) Serum total cortisol (␮g/dl) Refractory ascites and/or type-2 HRS (n, %)

31 (34%) 2 (2%) 39 (42%) 1 (1%) 3 (3%) 1 (1%) 2 (2%) 2 (2%) 2 (2%) 1 (1%) 9 (10%) 36 (39%) 28 (30%) 13 (14%) 20 (22%) 6 (6%) 9 (8–10); 50/43 19 (15–23) 7.3 (5.3–10.1) 24 (26%)

14 (42%) 0 (0%) 12 (37%) 0 (0%) 2 (6%) 0 (0%) 0 (0%) 1 (3%) 0 (0%) 0 (0%) 4 (12%) 7 (21%) 8 (24%) 4 (14%) 6 (18%) 2 (6%) 9 (8–11) 19 (15–24) 7.4 (5.5–10.4) 11 (33%)

2 (10%) 1 (5%) 14 (66%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 1 (5%) 3 (14%) 13 (62%) 8 (38%) 3 (14%) 5 (24%) 2 (10%) 9 (8–10) 19 (16–21) 7.3 (5.8–11.5) 4 (19%)

0.01 0.39 0.05

<0.01 0.27 0.85 0.66 0.66 0.56 0.41 0.97 0.25

Oesophageal varices No varices (n, %) F1 (n, %) F2 (n, %) F3 (n, %) Eradicated (n, %) Serum albumin (g/dl) Serum bilirubin (mg/dl)/(␮mol/l) INR (ratio) Mean arterial pressure (mmHg) Serum sodium (mmol/l) Serum creatinine (mg/dl)/(␮mol/l) Plasma renin activity (ng/ml/h) Serum aldosterone (pg/ml) Serum norepinephrine (pg/ml) Serum cholesterol (mg/dl) Serum HDL cholesterol (mg/dl) Serum A1-apolipoprotein (mg/dl) WBC (109 /l) CRP (mg/l)

19 (20%) 30 (33%) 27 (29%) 4 (4%) 13 (14%) 3.3 (3.0–3.6) 3.4 (2.0–6.5)/58 (34–111) 1.77 (1.53–2.13) 83 (76–93) 136 (132.5–138) 1.0 (0.7–1.4)/87 (63–124) 6.7 (3.4–14.4) 211 (98–339) 519 (302–857) 85 (57–102) 27 (14–36) 47 (21–77) 4375 (3310–6060) 10.55 (5.6–18.6)

8 (24%) 11 (33%) 10 (31%) 2 (6%) 2 (6%) 3.3 (3.0–3.7) 3.5 (1.5–7.4)/60 (26–127) 1.90 (1.60–2.32) 83 (77–93) 137 (135–139) 1.0 (0.7–1.4)/87 (63–124) 4.0 (1.4–11.0) 193 (98–314) 330 (234–589) 87 (56–104) 26 (14–33) 45 (22–76) 4220 (2670–6320) 13.6 (5.6–22.4)

4 (19%) 5 (24%) 6 (28%) 2 (10%) 4 (19%) 3.2 (3.0–3.5) 3.3 (2.3–5.1)/56 (39–87) 1.61 (1.54–1.89) 83 (77–89) 136 (130–138) 1.0 (0.8–1.2)/87 (70–106) 8.5 (5.4–19.7) 268 (126–557) 718 (374–1114) 85 (62–92) 29 (13–41) 56 (29–92) 4090 (3480–4520) 10.5 (5.7–18.6)

0.59 0.53 0.70 0.33 0.16 0.26 0.21 0.04 0.25 0.12 0.54 0.12 0.08 < 0.01 0.95 0.31 0.38 0.43 0.49

0.25 0.42 0.39 0.82

Continuous variables are expressed as median (interquartile range). Alcohol abuse was defined for a daily alcohol consumption >30 g in men and >20 g in women. HBV, hepatitis B virus; HCV, hepatitis C virus; HDV, hepatitis D virus; SBP, spontaneous bacterial peritonitis; HRS, hepatorenal syndrome; MELD, Model for End-stage Liver Disease; HDL, high density lipoproteins; WBC, white blood cells count; CRP, C-reactive protein.

survived and who did not undergo transplantation were censored at the last contact. In groups categorized according to the definitions of AD based on the cut points of basal, delta, and peak cortisol levels, the CID functions were estimated using the Kaplan–Meier product-limit method. The differences between groups were evaluated by Cox proportional-hazards models, adjusting for age, Model for End-Stage Liver Disease (MELD) score, and presence in the LT waiting list. The associations of basal, delta, and peak cortisol levels with mortality were also evaluated using their continuous values in the Cox proportional-hazards models. To explore the potential nonlinear relationships of these variables with mortality, the continuous measurements were included in the models as restricted cubic splines. The subgroup of patients evaluated for the presence of microbial DNA was also considered separately in the survival analysis, in order to estimate the effect of microbial DNA. In these patients, the association between the MELD score and serum cortisol level was assessed by the Pearson correlation coefficient (r) calculation. For the sensitivity analysis, comparison between microbial DNA positive and negative patients was also

adjusted using Cox models including a propensity score (PS) for the likelihood of the patient testing positive for microbial DNA. The PS was derived from a logistic model that included all the variables reported in Table 1, which in a stepwise forward selection showed a p of <0.30. In the survival analysis, an additional comparison between the groups was performed adjusting for PS and including the MELD score and presence in LT waiting list, which were not included in the PS.

3. Results During the study period, 207 consecutive inpatients with ascites were screened. Ninety-three (45%) patients admitted to the hospital for ascites and fulfilling the inclusion and exclusion criteria were enrolled; 114 patients were excluded due to infection (n = 65), bleeding (n = 19), multifocal hepatocellular carcinoma (n = 22), major surgery (n = 4), age (n = 2), and ongoing steroid treatment (n = 2).

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

4

Table 1 shows the baseline characteristics of the included patients. The median MELD score was 19 (interquartile range (IQR) 15–23), and the median Child–Pugh score was 9 (IQR 8–10, Child–Pugh B, n = 50, Child–Pugh C, n = 43). Twenty-four patients (26%) had ascites which was refractory to diuretic therapy, of whom 10 fulfilled the International Club of Ascites criteria for hepatorenal syndrome type 2 [26]. At the time of inclusion 54 patients were in the waiting list for LT (58%), 31 had been excluded because of contraindications (33%), and 8 were being assessed for LT (9%). The median value of basal serum cortisol was 7.3 ␮g/dl (IQR 5.3–10.1). The median values of PRA, serum aldosterone, and serum norepinephrine were 6.7 ng/ml/h (IQR 3.4–14.4), 211 pg/ml (IQR 98–339), and 519 pg/ml (IQR 302–857), respectively. The median follow-up duration was 369 days (range: 13–1273). Twenty one patients died (23%, 1-year mortality: 26.8%, 95% confidence interval (CI): 17.5–39.6) and 29 underwent LT (31%). The causes of death were liver failure in 10 (48%), type 1 hepatorenal syndrome in 5 (24%), sepsis and its complications in 4 (19%), and portal hypertension-related bleeding in 2 (9%).

Fig. 1. Cumulative (Kaplan–Meier) incidence of death in microbial DNA-positive (discontinuous line) and -negative (continuous line) patients.

3.1. Bacterial and fungal DNA The determination of bacterial and fungal DNA was performed in 54 serum samples, 54 heparinized blood samples, and 53 ascitic fluid samples from the 54 patients consecutively enrolled in the study after October 2009. We found no evidence of a different risk of mortality in this subgroup of patients compared to those for whom microbial DNA determinations were not carried out (1-year mortality: 74.9% vs. 71.9%, p = 0.40). Of the 54 patients, 8 were positive for bacterial DNA (15%, Staphylococcus aureus, n = 1; Sphingomonas spp., n = 1; Caldimonas spp., n = 1; coagulase negative Staphylococcus, n = 2; Staphylococcus coagulase negative Propionibacterium spp., n = 1; undetermined, n = 2) and 5 were positive for fungal DNA (9%, Epicoccum spp., n = 2; undetermined, n = 3). In one of the 54 patients, admitted to our unit because of hepatic encephalopathy and ascites, the ascitic fluid sample was not obtained as diagnostic paracentesis was unsuccessful. Three of 53 ascitic samples (6%) were positive for fungal DNA (Rhizopus spp., n = 1; Cladosporium spp., n = 1; and undetermined, n = 1) and 11 for bacterial DNA (21%). The following bacteria were detected: Escherichia coli (n = 2), Streptococcus spp. (n = 1), Staphylococcus epidermidis (n = 1), Corynebacterium spp. (n = 1), Pseudomonas spp.+Staphylococcus spp. (n = 1) and Staphylococcus epidermidis + Propionibacterium spp. (n = 1), and 4 undetermined cases. Overall, microbial DNA was detected in one or more fluids in 21 patients (39%), and in both blood and ascites in 5 (9%). These patients were not treated with antibiotics or antifungals, as they did not show any clinical features of overt infection. The baseline demographic, clinical, and laboratory characteristics of this subgroup of 54 patients are shown in Table 1. The use of treatments that may affect microbial DNA (such as antibiotic prophylaxis for recurrence of spontaneous bacterial peritonitis and/or hepatic encephalopathy, proton-pump inhibitors and betablockers for prevention of portal hypertension-related bleeding) did not differ significantly between microbial DNA positive and negative patients (p = 0.10, p = 0.90, and p = 0.50, respectively). 3.2. AD and mortality The prevalence of AD differed considerably depending on the criterion used to define it: 75% (70/93 patients) according to “basal cortisol,” 15% (14/93) according to “delta cortisol,” and 32% (30/93)

according to “peak cortisol.” The relationship between the presence of AD according to each criterion and mortality was weak: it showed a statistical significance at univariable analysis only using the criterion “delta cortisol” (hazard ratio (HR) 2.92; 95% CI 1.18–7.25), and, above all, no criteria were demonstrated to be significantly related to mortality on multivariable analysis. We then analyzed the relationship between mortality hazard and the basal, delta, and peak cortisol levels considered as continuous variables, in order to determine whether these indicators of adrenal function were predictive of mortality, irrespective of the generally used cutoff values. As no strong evidence of a threshold effect was detected, in the Cox models, we assumed that the relationship is linear for all cortisol determinations. Table 2 shows the results of the univariable and multivariable Cox models for survival. The mortality hazard increases with the rise in the basal cortisol values. This association was statistically significant at both univariable (HR 1.13 per 1-␮l/dl increase; 95% CI 1.02–1.26) and multivariable (HR 1.13 per 1-␮l/dl increase; 95% CI 1.01–1.26) analyses. No significant association was found between mortality and delta or peak cortisol level (at multivariable analysis, delta cortisol: HR 1.09 per 1-␮l/dl decrease, 95% CI: 0.99–1.20; peak cortisol: HR 1.01 per 1-␮l/dl increase, 95% CI: 0.93–1.09).

3.3. Adrenal function, microbial DNA, and mortality 3.3.1. Mortality Fig. 1 shows the cumulative incidence of mortality by the presence or absence of microbial DNA in serum/blood and ascites samples. Patients positive for bacterial/fungal DNA had a significantly (HR 5.81, 95% CI: 1.48–22.9; p = 0.01) lower survival (1-year mortality: 49.1%, 95% CI: 26.4–77.4) compared to those with negative microbiological results (1-year mortality: 11.0%, 95% CI: 3.6–30.6). At multivariable Cox analysis, only MELD score (HR ranging from 1.22 to 1.24 per unit increase) and presence of microbial DNA (HR ranging from 7.29 to 8.05) were found to have a statistically significant association with mortality, independently from the cortisol variable included (Table 3). The sensitivity analysis showed an increased risk of mortality in patients positive for microbial DNA, even adjusting the comparisons for PS: HR = 6.38, p = 0.027 adjusting only for PS; and HR = 8.05, p = 0.012 adjusting for PS, MELD score, and presence in the LT waiting list.

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

5

Table 2 Relationship between values of basal, delta and peak cortisol, considered as continuous variables, and mortality in the total cohort (93 patients). Cox proportional-hazard models. Presence of AD according to different diagnostic criteria Basal HR (95% CI)

p

Delta HR (95% CI)

p

Peak HR (95% CI)

p

Univariable analysis Cortisol valuea

1.13 (1.02–1.26)

0.018

1.13 (1.02–1.24)

0.014

1.03 (0.95–1.11)

0.467

Multivariable analysis Cortisol valuea MELD score (per 1 point increase) LT WL (yes vs. no) Age (per 1 year increase)

1.13 (1.01–1.26) 1.18 (1.09–1.27) 0.64 (0.20–2.05) 1.03 (0.98–1.09)

0.035 <0.001 0.453 0.272

1.09 (0.99–1.20) 1.15 (1.07–1.24) 0.61 (0.19–1.97) 1.04 (0.98–1.09)

0.098 <0.001 0.412 0.185

1.01 (0.93–1.09) 1.17 (1.08–1.26) 0.61 (0.19–1.95) 1.05 (0.99–1.10)

0.859 <0.001 0.404 0.100

AD, adrenal dysfunction; MELD, Model for End-stage Liver Disease; LT WL, liver transplantation waiting list. a The hazard ratios refer to a 1 ␮g/dl increase for basal and peak, and to a 1 ␮g/dl decrease for delta.

Table 3 Relationship between values of basal, delta and peak cortisol, considered as continuous variables, and mortality in 54 patients with microbial DNA test. Cox proportional-hazard models. Presence of AD according to different diagnostic criteria Basal HR (95% CI)

p

Delta HR (95% CI)

p

Peak HR (95% CI)

p

Univariable analysis Cortisol valuea

1.08 (0.91–1.29)

0.379

1.10 (0.97–1.25)

0.119

1.05 (0.93–1.18)

0.419

Multivariable analysis Cortisol valuea MELD score (per 1 point increase) LT WL (yes vs. no) Microbial DNA* (present vs. absent)

1.00 (0.80–1.24) 1.24 (1.07–1.42) 0.61 (0.15–2.56) 8.05 (1.57–41.24)

0.990 0.004 0.501 0.012

1.05 (0.93–1.19) 1.22 (1.07–1.40) 0.64 (0.18–2.32) 7.29 (1.47–36.17)

0.392 0.004 0.497 0.015

1.05 (0.93–1.17) 1.24 (1.08–1.41) 0.55 (0.15–1.99) 7.94 (1.62–38.88)

0.434 0.002 0.366 0.011

AD, adrenal dysfunction; MELD, Model for End-stage Liver Disease; LT WL, liver transplantation waiting list. a The hazard ratios refer to a 1 ␮g/dl increase for basal and peak, and to a 1 ␮g/dl decrease for delta. * Univariable HR 5.81 (95% CI: 1.48–22.85, p = 0.012).

3.3.2. Microbial DNA and adrenal function The relationship between MELD score values and the different serum cortisol levels was explored, stratifying patients according to positivity or negativity for microbial DNA in organic fluids. A significant association between the basal serum cortisol level and MELD score values was found only in patients positive for the presence of microbial DNA (Pearson’s r = 0.5107; p = 0.0180), while the delta and peak cortisol levels did not show any relevant associations with

MELD score values in both microbial DNA negative and positive patients (Fig. 2).

4. Discussion The syndrome named “critical-illness-related corticoid insufficiency” is well known as a serious complication in severely ill

Fig. 2. Correlations between the basal cortisol and MELD score values in all patients (A), microbial DNA-negative patients (B), and microbial DNA-positive patients (C). MELD, Model for End-Stage Liver Disease.

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7 6

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

patients [1–4]. Surprisingly, little is known of adrenal function in liver cirrhosis, especially in noncritical patients [12–16,27]. In this study, we investigated this issue, focusing on noninfected haemodynamically stable patients. Using the three different criteria of AD widely accepted in the literature, neither the prevalence of AD nor its prospective evaluation as a predictor of mortality yielded consistent results. Our data highlight the poor concordance of the traditional criteria in defining AD and their lack of independent prognostic power in this clinical setting. These results are only apparently in contrast with those published in a recent study [12], because >60% of those patients were studied either when infected or <24 h after a variceal bleed, two well-known causes of adrenal stimulation. This makes our population hardly comparable with the patients in the abovementioned study. Our second step was the evaluation of the relationship between adrenal function itself and mortality, irrespective of the conventional AD criteria and cutoff points. Interestingly, data showed that the mortality risk increased in parallel with the increase in basal cortisol values. This result appeared to be very reliable, as confirmed by the high stability of the HR at multivariable analysis, even after the adjustment for a very strong mortality-predicting variable such as the MELD score. It is interesting to note whether similar results would have been acquired if the free cortisol levels were measured, which might estimate adrenal function more accurately. In fact, a decrease in the serum total cortisol level may reflect a reduction in the levels of cortisol-binding proteins and serum albumin (typical of advanced cirrhosis), rather than a decrease in the serum free cortisol level [14,25]. However, only in 7 patients (8%), the serum albumin concentration was <2.5 g/dl: above this cutoff value, serum albumin levels were reported to not significantly affect the assessment of adrenal function in cirrhotic patients [27]. Moreover, the diagnostic cutoffs for the free cortisol values have not been adequately defined yet [13], and their determination is not routinely used, because of the complexity of the technique and high costs. Nevertheless, the findings described above indicate that adrenal activation with increased hormonal output is associated with an increased risk of mortality in noncritically ill patients with advanced cirrhosis. An explanation for this adrenal activation may be an underlying chronic inflammatory state, caused by microbial translocation from the gut lumen to the organism promoted by portal hypertension [19,20]. Therefore, the presence of microbial DNA was determined in the last 54 consecutive cirrhotic patients enrolled in our study, and it correlated with adrenal function and mortality. The data confirmed the significant sound association between microbial DNA and a poor prognosis [21]. Both the MELD score and microbial DNA were confirmed as effective and independent predictors of mortality, while adrenal activation did not maintain its prognostic value when microbial DNA was introduced into the statistical models. The question of whether adrenal function would still be relevant for the clinical outcome if studies investigating adrenal function in non-septic cirrhotic patients had assessed the presence of microbial DNA is intriguing. Clearly, we do not have the answer, but the results of our survival analyses changed dramatically when microbial DNA was introduced in the model; at the same time, adrenal function was found to be less relevant, perhaps being a mere epiphenomenon of the presence of an occult infection. Furthermore, a clear association between liver function deterioration (measured by MELD score) and adrenal activation emerged in patients positive for microbial DNA. It is plausible that, in this context, bacterial and fungal DNA activate the immune host response inducing the release of pro-inflammatory cytokines, thereby contributing to the worsening of liver function. As microbial DNA toxins and the secreted cytokines are also known to

activate the hypothalamic–pituitary–adrenal axis [22], the adrenal stimulation observed in our patients goes in parallel with liver deterioration and it reflects the glandular response to a clinically silent status of infection, even in the case of non-strictly pathogenic microorganisms. It is interesting to note that microbial DNA positive patients had a significantly higher previous hepatic encephalopathy rate and that their neurohormonal endogenous systems were more activated compared to those negative for microbial DNA, reflecting the clinical and haemodynamic relevance of microbial translocation in cirrhotic patients with ascites (see Table 1). It should be noted that the important findings of this study are based on the subgroup of 54 consecutive patients who underwent the microbial DNA test. The limited sample size represents the major limitation of our results and a strong reason for promoting independent validations. In conclusion, the results of this study showed that adrenal function was a weak predictor of death in noninfected haemodynamically stable cirrhotic patients with ascites, and that only the presence of microbial DNA and the MELD score were independent predictors of mortality. For the first time, our data indicate that there is a direct association between the degree of liver insufficiency and the activation of adrenal function in patients testing positive for microbial DNA; high cortisol levels may reflect the attempt of the host to respond to the subclinical infectious status. If these findings are confirmed, then a new prognostic score could be developed, which also considers the variable of fluid contamination from bacterial and fungal DNA, in order to better identify patients with a worse prognosis. Conflict of interest None declared. References [1] Marik PE, Pastores SM, 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. Critical Care Medicine 2008;36:1937–49. [2] Annane D, Maxime V, Ibrahim F, et al. Diagnosis of adrenal insufficiency in severe sepsis and septic shock. American Journal of Respiratory and Critical Care Medicine 2006;174:1319–26. [3] Cooper MS, Stewart PM. Corticosteroid insufficiency in acutely ill patients. New England Journal of Medicine 2003;348:727–34. [4] Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. Journal of American Medical Association 2002;288:862–71. [5] Annane D, Bellissant E, Boallert PE, et al. Corticosteroids for severe sepsis and septic shock: a systematic review and meta-analysis. British Medical Journal 2004;329:480. [6] Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock. New England Journal of Medicine 2008;358:111–24. [7] Tsai MH, Peng YS, Chen YC, et al. Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock. Hepatology 2006;43:673–81. [8] Harry R, Auzinger G, Wendon J. The clinical importance of adrenal insufficiency in acute hepatic dysfunction. Hepatology 2002;36:395–402. [9] Fernandez 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;44:1288–95. [10] Marik PE, Gayowski T, Starzl TE. The hepatoadrenal syndrome: a common yet unrecognized clinical condition. Critical Care Medicine 2005;33:1254–9. [11] Arabi YM, Aljumah A, Dabbagh O, et al. Low dose hydrocortisone in patients with cirrhosis and septic shock: a randomized controlled trial. Canadian Medical Association Journal 2010;182:1971–7. [12] Acevedo J, Fernandez 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, http://dx.doi.org/10.1002/hep.26535. [13] Fede G, Spadaro L, Tomaselli T, et al. Adrenocortical dysfunction in liver disease: a systematic review. Hepatology 2012;55:1282–91. [14] Fede G, Spadaro L, Tomaselli T, et al. Assessment of adrenocortical reserve in stable patients with cirrhosis. Journal of Hepatology 2011;54:243–50. [15] Trifan A, Chiriac S, Stanciu C. Update on adrenal insufficiency in patients with liver cirrhosis. World Journal of Gastroenterology 2013;19:445–56. [16] Acevedo J, Fernandez J. New determinants of prognosis in bacterial infections in cirrhosis. World Journal of Gastroenterology 2014;20:7252–9.

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010

G Model YDLD-2877; No. of Pages 7

ARTICLE IN PRESS A. Risso et al. / Digestive and Liver Disease xxx (2015) xxx–xxx

[17] Thevenot T, Borot S, Remy-Martin A, et al. Assessing adrenal function in cirrhotic patients: is there a reliable test. Gastroenterologie Clinique et Biologique 2009;33:584–8. [18] Ruiz-del-Arbol L, Urman J, Fernandez J, et al. Systemic, renal and hepatic hemodynamic derangement in cirrhotic patients with spontaneous bacterial peritonitis. Hepatology 2003;38:1210–8. [19] Francés R, Munaz C, Zapater P, et al. Bacterial DNA activates cell mediated immune response and nitric oxide overproduction in peritoneal macrophages from patients with cirrhosis and ascites. Gut 2004;53:860–4. [20] Guarner C, Soriano G. Bacterial translocation and its consequences in patients with cirrhosis. European Journal of Gastroenterology & Hepatology 2005;17:27–31. [21] Zapater P, Francés R, Gonzalez-Navajas JM, et al. Serum and ascitic fluid bacterial DNA: a new independent prognostic factor in noninfected patients with cirrhosis. Hepatology 2008;48:1924–31. [22] Bellot P, Garcia-Pagan JC, Francés R, et al. Bacterial DNA translocation is associated with systemic circulatory abnormalities and intrahepatic endothelial dysfunction in patients with cirrhosis. Hepatology 2010;52:2044–52.

7

[23] Nouh A, Elhendy A, Eldeeb G. Spontaneous fungal peritonitis: occurrence in patients with cirrhotic ascites. Menoufiya Medical Journal 2008;2: 5–16. [24] Crowley S, Hindmarsh PC, Honour JW, et al. Reproducibility of the cortisol response to stimulation with a low dose of ACTH (1–24): the effects of basal cortisol levels and comparison of low dose with high dose secretory dynamics. Journal of Endocrinology 1993;136:167–72. [25] Giordano R, Picu A, Bonelli L, et al. Hypothalamus-pituitary-adrenal axis evaluation in patients with hypothalamo-pituitary disorders: comparison of different provocative tests. Clinical Endocrinology 2008;68:935–41. [26] Moore KP, Wong F, Gines P, et al. The management of ascites in cirrhosis: report on the consensus conference of the International Ascites Club. Hepatology 2003;38:258–66. [27] 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 International 2011;31:425–33.

Please cite this article in press as: Risso A, et al. Adrenal function and microbial DNA in noninfected cirrhotic patients with ascites: Relationship and effect on survival. Dig Liver Dis (2015), http://dx.doi.org/10.1016/j.dld.2015.04.010