Cirrhosis constitutes the ultimate stage of fibrosing liver diseases and is associated with complications leading to hospitalization and death (1). The WHO recently estimated cirrhosis to be the 12th cause of mortality in the world, with a number of associated deaths greater than 1 million per year (2). Acutely ill cirrhotic patients often require admission to the ICU (3, 4). This subgroup of patients has a poor outcome in the ICU compared to the general critically ill population; recent studies report that the death rate ranges from 18% to 66% when all reasons of admission are considered (5–10).
Sepsis and septic shock frequently complicate cirrhosis and are among the main reasons for admission to the ICU for these patients (3, 7, 9, 10). Indeed, cirrhosis associates an immunosuppressed state and complex organ alterations that often lead to multiple organ failures during sepsis. Data regarding cirrhotic patients with septic shock are scarce (11–14). Organ supports are often required in these patients and have been shown as strong predictors of short-term mortality (13). However, little is known about long-term outcome.
Thus, we conducted a long-term observational study of a cohort of cirrhotic patients admitted to the ICU for septic shock in order to describe outcomes and to analyze potential risk factors for both short and long-term mortality.
We undertook a retrospective analysis of a prospective cohort investigation of all adult cirrhotic patients with septic shock admitted to a 15-bed university-affiliated medical ICU from January 2002 to December 2013. The hospital hosts a gastroenterology and hepatology department. The study received approval from the local ethics committee (Comité de Protection des Personnes Lyon Sud-Est II). This institutional review board waived the need for informed consent in view of the observational nature of the study.
Patients who were 18 years or older, who had cirrhosis, and who were admitted to the ICU for septic shock were included. Patients were considered only if a diagnosis of cirrhosis was established before or during their stay in the ICU. Cirrhosis was diagnosed by a specialist in hepatology, based on available clinical, biological, and pathological data. Patients with prior liver transplantation were excluded. No extracorporeal liver support was used in the ICU. Adapted from the Surviving Sepsis Campaign, septic shock was defined by the ICU physician in charge of the patient, in the context of probable or documented infection, as persistently low mean arterial pressure (i.e., <65 mmHg) that required vasopressors despite initial fluid resuscitation (15). All diagnoses of septic shock were retrospectively confirmed by two independent senior intensivists. Patients were managed according to applicable guidelines (15).
Data regarding demographics and prior clinical status were collected: age, sex, underlying diseases and Charlson score, functional status according to the Knaus classification, McCabe scale, etiology of cirrhosis, Child–Pugh score (CPS) at baseline, complications of cirrhosis, and assessment for liver transplantation (16–19). The following data obtained at admission were collected: clinical signs, CPS, Model for End-stage Liver Disease score, Sequential Organ Failure Assessment (SOFA), organ supports, and Simplified Acute Physiology Score II (20–22). Microbiological data from admission samples, diagnosis of infected sites, and antibiotic treatments were also collected. Outcomes in the ICU and at 1 year were registered.
CPS reported by the attending hepatologist the year before admission to the ICU defined the CPS at baseline. When cirrhosis was unknown before admission to ICU, CPS class was retrospectively computed with available data for the year before admission. Data collected at admission was the worst value during the first 24 h in ICU. Encephalopathy was classified according to West Haven criteria (23). As previously described, for each of the six vital functions (i.e., cardiovascular, respiratory, neurologic, renal, liver, and hematologic), a SOFA sub-score >2 defined organ failure (24). Diagnosis of Acute Respiratory Distress Syndrome (ARDS) was made in accordance with the 1994 American–European consensus conference (25). Infection was reported as nosocomial when related symptoms had begun 48 h or more after hospital admission. For culture-positive infections, initial antibiotic therapy was defined appropriate when the identified infectious agents were known to be regularly susceptible to at least one of the drug included in the regimen.
Data were expressed as number (%) or mean (SD), as appropriate. In order to evaluate the potential influence of time, patient characteristics, and outcomes were compared among three balanced periods: 2002–2005, 2006–2009, and 2010–2013. Risk factors for both ICU and 1-year mortalities were assessed. For this, univariate analyses were first performed using Fisher's exact test for categorical variables or Mann–Whitney U test for quantitative variables. All variables with P value <0.10, as well as age and sex, were entered in a backward stepwise multivariate analysis using a logistic regression model. CPS on admission was split into components (i.e., prothrombin time, ascites, bilirubin, encephalopathy, and albumin). We chose to assess the influence of organ supports (e.g., renal replacement therapy) rather than related organ failures (e.g., renal failure) or SOFA sub-scores, which were both excluded from the model to avoid redundancy. Potential confounding factors were eliminated if P value was >0.15. Odds ratios (OR) were estimated with 95% confidence interval (95% CI). The goodness-of-fit of the model was evaluated using the Hosmer–Lemeshow test.
Survival curves were constructed with the Kaplan–Meier method according to the independent factors predicting 1-year mortality. Survival distributions were compared by use of the Log-rank test.
All statistical tests were two-tailed and differences were considered statistically significant with P value <0.05. Data were analyzed using SPSS software version 19.0 (IBM Corp. Released 2010. Armonk, NY).
During the study period, 7,644 patients were admitted in the ICU among whom 149 (2%) were cirrhotic patients with septic shock. Patients were admitted from the emergency ward or directly from home in 54 (36%) cases, from a medical ward in 41 (28%) cases, and from an intermediate care unit in 54 (36%) cases. Main data regarding demographics, comorbidities, characteristics of cirrhosis, ICU, and 1-year mortalities were not significantly different across the predefined periods (See Table, Supplemental Digital Content 1, https://links.lww.com/SHK/A828).
The main characteristics of patients are reported in Table 1. Sex ratio was 2.4 and 79 (53%) patients were aged 60 years or older; the maximal age was 85 years. The most frequent cause of cirrhosis was abusive alcohol consumption (Table 1). A total of 106 (71%) patients acknowledged current excessive alcohol consumption at admission. Twenty-one (14%) patients had previously been admitted to an ICU for a cirrhosis-related reason.
Clinical and biological data at ICU admission are presented in Table 2. The site of causal infection was identified in 117 (79%) patients; pneumonia, the most frequent etiology, was found in 58 (39%) patients. Infection was nosocomial in 52 (35%) cases. Microbiological samples yielded contributive results in 110 (74%) cases. Blood cultures were positive in 64 (43%) cases including 13 (9%) patients for whom the site of infection remained unknown after diagnostic work-up. Among 95 monomicrobial infections (irrespective of sample site), gram-positive cocci were identified in 38 (40%) cases and gram-negative rods in 45 (30%) cases. Multidrug-resistant bacteria were found in 18 (1%) cases. Fungal agents were isolated as causative pathogens in 12 (8%) patients, including Candida species in 10 (7%) cases. As expected, all patients required norepinephrine and antibiotic treatment after admission to the ICU. Among documented infections, initial antibiotic therapy was appropriate in 85% of cases (94/110). During the first 24 h, 100 (67%) patients had three or more organ failures and 111 (74%) had a SOFA score ≥12.
Length of ICU stay was 10 ± 11 days. End-of-life decisions occurred in 26 (17%) patients. Sixty-nine (46%) patients were discharged alive from the ICU. Regarding management of organ failure over the course of ICU stay, duration of norepinephrine use was 5 ± 6 days and mechanical ventilation (duration: 9 ± 10 days) was needed in 125 (84%) patients including 44 (30%) cases that fulfilled ARDS criteria. Renal replacement therapy was initiated in 60 (40%) patients in ICU.
Among admission variables identified as susceptible to influence short-term outcome by univariate analysis (Table 2), the following were found to be independently associated with ICU mortality in multivariate analysis: renal replacement therapy (OR 13.95, 95% CI 3.30;59.03), arterial lactate >5 mmol.L−1 (OR 7.27, 95% CI 2.92;18.10), and mechanical ventilation (OR 3.05, 95% CI 1.08;8.58). The Hosmer–Lemeshow test validated the goodness-of-fit of the model (P = 0.81).
Only 40 (27%) patients were alive at 1 year after ICU admission; among the 69 ICU survivors, 29 (42%) died after ICU discharge. No patient underwent liver transplantation during follow-up; 3 additional patients were listed during this period.
Following univariate analysis (data not shown), the logistic regression model found that the need for renal replacement therapy during the ICU stay was the strongest predictor for 1-year mortality (Table 3). Furthermore, prothrombin time ≤40% on admission and Charlson score, which are linked to hepatic condition and underlying diseases, were also independently associated with long-term outcome. However, mechanical ventilation was not significantly associated with 1-year mortality (Table 3). Goodness-of-fit of the model was validated with the Hosmer–Lemeshow test (P = 0.20).
Kaplan–Meier survival curves according to predictors of 1-year mortality are presented in Figure 1. Notably, 1-year survival rate was only 7% (4/60) in patients who received renal replacement therapy during the ICU stay versus 40% (36/89) in those who did not.
In the present study, we report for the first time high rates of both ICU and 1-year mortality in a homogenous population of cirrhotic patients admitted to ICU with septic shock. Short-term outcome was independently predicted by the initial arterial lactate level, and the early need for either mechanical ventilation or renal replacement therapy. The weight of the underlying liver disease appeared to impact mainly the long-term outcome. The need for renal replacement therapy in the ICU was also associated with an extremely poor survival rate at 1 year.
The poor prognosis of critically ill cirrhotic patients, when compared to noncirrhotic ones in the ICU, has been recognized for a long time (26). However, ICU mortality rate of cirrhotic patients varies widely in recent studies and determinants of the outcome remain unclearly defined (5–14). This is partly due to the heterogeneity of the ICU population selected only on the basis of cirrhosis. Indeed, various events can precipitate the organ failure leading to the ICU admission (5–10). Furthermore, cirrhotic patients may either be admitted with acute or chronic liver failure or not (3, 5–10). Interestingly, the reason for ICU admission has been shown to influence short-term outcome in a few studies (7, 10). In particular, most of the recent literature agrees that sepsis, which is involved in 15% to 41% of admissions, is an independent prognostic factor in these patients (3, 5–7, 9, 10, 27). In the present work, we chose to focus on septic shock, the most severe form of sepsis.
The increased susceptibility of cirrhotic patients to severe infections is not fully understood. It notably involves a systemically impaired immunity with deregulated proinflammatory reactions (28, 29). As also reported herein, usual infection sites, such as pneumonia and urinary tract infections, are among the most frequent infections described in cirrhotic patients (27, 28). Furthermore, almost 40% of the patients herein presented with either spontaneous bacterial peritonitis or infection of unknown origin, which is often related to intestinal microbial translocation. Indeed, digestive translocation is a major determinant of severe infectious diseases in cirrhotic patients; first, it directly causes clinical infections, i.e., peritoneal or without focus (27, 30), and second, even when not the source of sepsis, intestinal microbial translocation acts as an aggravating factor, inducing vasodilatation that further deteriorates hemodynamics (27, 31, 32). The impact of the source of infection on ICU mortality has been questioned in cirrhotic patients with septic shock (12, 13). In the present study, neither the site of infection nor the baseline characteristics were predictive of short-term outcome.
Concerning the outcomes of septic shock for cirrhotic patients, studies covering the last decade showed ICU death rates ranging from 60% to 80% (11–14). The 54% ICU mortality rate herein is below the lower bound of this range. We did not observe any change in survival during the study period, yet two previous studies have described a trend toward improved ICU outcome since the late 1990s in France (12, 13). Analyzing data from a French ICU network, Galbois et al.(13) reported an ICU mortality decline since the early 2000s. In the same way, Sauneuf et al.(12) studied two consecutive periods, reporting a significant decrease in ICU mortality over time. Although a few characteristics of the patients from these studies (e.g., demographics, comorbidities, or organ supports) appeared similar to those of the present study population, further comparison raises several concerns. First, at an equivalent level of illness severity in the two periods considered, the extremely high mortality rate observed during the period going from the late 1990s to early 2000s remains unexplained in these studies (12, 13). Second, comparison of organ failure and liver function scores is limited as these data were not available in the largest study (13). Third, improvements in management of sepsis and/or cirrhosis could have been heterogeneously implemented over time, with potential influence on outcomes. Finally, the lack of data regarding end-of-life decisions, which were not systematically reported, greatly impedes comparison of outcomes.
This cohort provides important data regarding the prognostic evaluation for cirrhotic patients with septic shock. Only a few recent papers have assessed outcome predictors for this specific ICU population (11–13). Yet, prognostication of patients at ICU admission is of crucial importance to guide improvements in both quality of care and allocation of resources. It has been previously demonstrated that organ failure score or organ supports are prognostic factors of unfavorable outcome (12, 13). In the present study, early use of additional organ supports other than vasopressors, i.e., mechanical ventilation and/or renal replacement therapy, was strongly associated with ICU mortality. In addition, a lactate value on admission >5 mmol.L−1 was independently associated with worse short-term outcome. Hyperlactatemia was not always proved to be an independent prognostic factor in this setting (11, 12). However, this finding is of some importance because it could reflect the magnitude of the systemic hypoperfusion preceding the ICU stay. Furthermore, it could also be a surrogate for liver failure, suggesting an impaired lactate clearance in the patients herein (33). Regarding liver severity scores, some debate remains as to their value as outcome predictors in ICU (4). Sauneuf et al.(12) previously reported that usual application of the CPS, i.e., before the ICU admission, could better predict short-term prognosis than CPS used as a score of acute liver failure, i.e., calculated using admission data. In the present study, neither the CPS class the year before admission, nor the CPS at admission were associated with ICU mortality. We acknowledge as a limitation that we did not use other recent scores (e.g., the Chronic Liver Failure–SOFA, CLIF–SOFA), which might also contribute to the outcome prediction in critically ill cirrhotic patients (34). Nevertheless, the severity of liver failure on admission, as represented by a prothrombin time ≤40%, was independently associated with long-term outcome. Overall, a key message of the study remains the high 1-year mortality rate observed. To the best of our knowledge, this is the first study reporting 1-year mortality in the septic shock population of cirrhotic patients. Thus, roughly three out of four patients had died at 1 year, with a dramatic death rate of 42% in ICU survivors. The results emphasize also that renal replacement therapy, when required during the ICU stay, heavily altered the long-term outcome with a 93% 1-year mortality rate.
This study does, however, have some limitations. First, and although data were prospectively collected through the patient data management system, we acknowledge that the retrospective design of the study allows only hypotheses to be made regarding risk factors of mortality for cirrhotic patients with septic shock. Second, the study was conducted in a single center, hindering the generalization of the results. On the other hand, and in view of the scarcity of the disease, we report herein one of the largest cohorts of cirrhotic patients with septic shock admitted to ICU. Third, we did not use the last definition for septic shock (i.e., Sepsis-3) because of the period covered by the study (35). Nevertheless, a large majority of patients fulfilled the criteria of this new definition. Even if medical records were thoroughly reviewed by independent physicians, we cannot totally exclude few overdiagnosis of septic shock due to the well-admitted uncertainty in the definition of infection. Fourth, long-term quality-of-life data were not available. Further studies with adequate post-ICU follow-up are needed to allow a comprehensive understanding of long-term outcome determinants in these patients.
In summary, cirrhotic patients with septic shock have a high ICU mortality rate, which is strongly predicted by initial illness severity and organ support requirement. Whereas underlying hepatic disease seems not to impact short-term prognosis, it might influence 1-year outcome. When renal replacement therapy is needed in the ICU, 1-year survival rate dramatically decreases.
We thank Dr Philip Robinson (DRCI, Hospices Civils de Lyon) for help in manuscript preparation.
1. Tsochatzis EA, Bosch J, Burroughs AK. Liver cirrhosis. Lancet
2. World Health Organization: Global burden of disease. http://www.who.int/healthinfo/global_burden_disease/estimates/en/index1.html
. Accessed September 28, 2018.
3. Olson JC, Kamath PS. Acute-on-chronic liver failure: concept, natural history, and prognosis. Curr Opin Crit Care
4. Saliba F, Ichaï P, Levesque E, Samuel D. Cirrhotic patients in the ICU: prognostic markers and outcome. Curr Opin Crit Care
5. Das V, Boelle PY, Galbois A, Guidet B, Maury E, Carbonell N, Moreau R, Offenstadt G. Cirrhotic patients in the medical intensive care unit: early prognosis and long-term survival. Crit Care Med
6. O’Brien AJ, Welch CA, Singer M, Harrison DA. Prevalence and outcome of cirrhosis patients admitted to UK intensive care: a comparison against dialysis-dependent chronic renal failure patients. Intensive Care Med
7. Levesque E, Hoti E, Azoulay D, Ichaï P, Habouchi H, Castaing D, Samuel D, Saliba F. Prospective evaluation of the prognostic scores for cirrhotic patients admitted to an intensive care unit. J Hepatol
8. Boone MD, Celi LA, Ho BG, Pencina M, Curry MP, Lior Y, Talmor D, Novack V. Model for End-Stage Liver Disease score predicts mortality in critically ill cirrhotic patients. J Crit Care
9. Levesque E, Saliba F, Ichaï P, Samuel D. Outcome of patients with cirrhosis requiring mechanical ventilation in ICU. J Hepatol
10. Theocharidou E, Pieri G, Mohammad AO, Cheung M, Cholongitas E, Agarwal B, Burroughs AK. The Royal Free Hospital Score: a calibrated prognostic model for patients with cirrhosis admitted to intensive care unit. Comparison with current models and CLIF-SOFA Score. Am J Gastroenterol
11. Arabi YM, Dara SI, Memish Z, Al Abdulkareem A, Tamim HM, Al-Shirawi N, Parrillo JE, Dodek P, Lapinsky S, Feinstein D, et al. Antimicrobial therapeutic determinants of outcomes from septic shock
among patients with cirrhosis. Hepatology
12. Sauneuf B, Champigneulle B, Soummer A, Mongardon N, Charpentier J, Cariou A, Chiche JD, Mallet V, Mira JP, Pène F. Increased survival of cirrhotic patients with septic shock
. Crit Care
13. Galbois A, Aegerter P, Martel-Samb P, Housset C, Thabut D, Offenstadt G, Ait-Oufella H, Maury E, Guidet B. Collège des Utilisateurs des Bases des données en Réanimation (CUB-Réa) Group:. Improved prognosis of septic shock
in patients with cirrhosis: a multicenter study. Crit Care Med
14. Galbois A, Bigé N, Pichereau C, Boëlle PY, Baudel JL, Bourcier S, Maury E, Guidet B, Ait-Oufella H. Exploration of skin perfusion in cirrhotic patients with septic shock
. J Hepatol
15. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, Gea-Banacloche J, Keh D, Marshall JC, Parker MM. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock
. Crit Care Med
16. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis
17. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE–acute physiology and chronic health evaluation: a physiologically based classification system. Crit Care Med
18. McCabe WR, Jackson GG. Gram-negative bacteremia: etiology and ecology. Arch Intern Med
19. Pugh RN, Murray–Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of oesophagus for bleeding oesophageal varices. Br J Surg
20. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, D’Amico G, Dickson ER, Kim WR. A model to predict survival in patients with end-stage liver disease. Hepatology
21. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (Sepsis related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-related Problems of the European Society of Intensive Care Medicine. Intensive Care Med
22. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on European/North American multicenter study. JAMA
23. Ferenci P, Lockwood A, Mullen K, Tarter R, Weissenborn K, Blei AT. Hepatic encephalopathy-definition, nomenclature, diagnosis, and quantification: Final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998. Hepatology
24. Vincent JL, De Mendonça A, Cantraine F, Moreno R, Takala J, Suter PM, Sprung CL, Colardyn F, Blecher S. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working Group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med
25. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M, LeGall JR, Morris A, Spragg R. Report of the American–European consensus conference on ARDS: definitions, mechanism, relevant outcomes and clinical trial coordination. The consensus committee. Intensive Care Med
26. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med
27. Wong F, Bernardi M, Balk R, Christman B, Moreau R, Garcia-Tsao G, Patch D, Soriano G, Hoefs J, Navasa M, et al. Sepsis in cirrhosis: report on the 7th meeting of the International Ascites Club. Gut
28. Gustot T, Durand F, Lebrec D, Vincent JL, Moreau R. Severe sepsis in cirrhosis. Hepatology
29. Deviere J, Content J, Denys C, Vandenbussche P, Schandene L, Wybran J, Dupont E. Excessive in vitro bacterial lipopolysaccharide-induced production of monokines in cirrhosis. Hepatology
30. Jalan R, Fernandez J, Wiest R, Schnabl B, Moreau R, Angeli P, Stadlbauer V, Gustot T, Bernardi M, Canton R, et al. Bacterial infections in cirrhosis: a position statement based on the EASL Special Conference 2013. J Hepatol
31. Lin RS, Lee FY, Lee SD, Tsai YT, Lin HC, Lu RH, Hsu WC, Huang CC, Wang SS, Lo KJ. Endotoxemia in patients with chronic liver diseases: relationship to severity of liver diseases, presence of esophaegeal varices, and hyperdynamic circulation. J Hepatol
32. Rasaratnam B, Kaye D, Jennings G, Dudley F, Chin-Dusting J. The effect of selective intestinal decontamination on the hyperdynamic circulatory state in cirrhosis. A randomized trial. Ann Intern Med
33. Drolz A, Horvatits T, Rutter K, Landahl F, Roedl K, Meersseman P, Wilmer A, Kluwe J, Lohse AW, Kluge S, et al. Lactate improves prediction of short-term mortality in critically ill cirrhosis patients: a multinational study. Hepatology
34. Lan P, Wang SJ, Shi QC, Fu Y, Xu QY, Chen T, Yu YX, Pan KH, Lin L, Zhou JC, et al. Comparison of the predictive value of scoring systems on the prognosis of cirrhotic patients with suspected infection. Medicine (Baltimore)
35. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, Bellomo R, Bernard GR8, Chiche JD9, Coopersmith CM, et al. The third international consensus definitions for sepsis and septic shock