Impact of chronic hepatitis C on mortality in cirrhotic patients admitted to intensive-care unit

Cirrhosis and mortality have been widely studied in individuals admitted to ICUs [10, 11, 15, 25, 26]; but to our knowledge, no previous studies have investigated the influence of CHC on ICU mortality among cirrhotic patients. Our manuscript provides a nationwide view of mortality rates of CHC-cirrhotic patients admitted into ICUs in Spain. In this study, the major findings were: 1) patients with CHC had a higher cumulative incidence of severe sepsis compared to patients without CHC in the subgroup of compensated cirrhosis, whereas the cumulative incidence of severe sepsis did not differ among patients with decompensated cirrhosis; 2) cirrhotic patients with CHC (CHC-group) had a greater tendency to die in ICU than non-CHC subjects (control-group); 3) The influence of CHC was observed on patients with compensated cirrhosis and severe sepsis, whereas CHC seemed to have no impact on mortality in patients with decompensated cirrhosis, independently of the onset of severe sepsis. Note that several baseline clinical characteristics showed statistically significant differences between groups, but these variables were included in the model to adjust the competitive risks.

There is controversial about whether patients with cirrhosis may benefit from the ICU management. In our study, the mortality was extremely high (merely 80 %) in comparison to data reported in the review of Saliba et al. [18]. There also are several important series published during the last 5 years with a mortality about 50 % [59 % Das et al. [27], 70.1 % Galbois et al. [28], 58.8 % Pan et al. [29], 68.1 % Bao et al. [30], 60 % McPhail et al. [31]]. These differences in death rates could be due to the designs of the studies used to obtain the data, which were very different, especially in our study carried out with data of Spanish MBDS from 2005 to 2010 and with a higher number of patients.

A systemic response to infection is more intense in the presence of cirrhosis, which translates into a greater risk of developing severe sepsis [14]. Thus, cirrhotic patients admitted into ICUs, apart from having a higher prevalence of infection than non-cirrhotic patients, have a higher rate of sepsis and death [13]. Furthermore, decompensated cirrhosis predisposes to delayed intestinal transit, increasing intestinal permeability and facilitating bacterial translocation from the gastro-intestinal lumen to the systemic circulation. This is accompanied by cirrhosis-associated immune dysfunction, which encourages systemic inflammation [32]. In our study, cirrhotic patients with and without CHC had similar rates of severe sepsis. However, when patients were stratified by the presence or absence of decompensated cirrhosis, the CHC-group with compensated cirrhosis had a higher cumulative incidence of severe sepsis than expected; which suggests that, in patients with compensated cirrhosis, susceptibility to severe sepsis may be increased by CHC.

The severity of infection is higher in cirrhotic patients than in non-cirrhotic patients [12, 15]. In addition, infections are increasingly recognized as a major trigger of systemic inflammation and organ failure in decompensated cirrhosis, leading to four-fold increased mortality rates [33]. In this setting, bacterial infections and sepsis are recognized as a distinct stage in the natural progression of chronic liver disease as they accelerate organ failure and contribute to the high mortality observed in decompensated cirrhosis [32]. In our study, as expected, cirrhotic patients with severe sepsis had a higher risk of ICU mortality than patients without severe sepsis; and patients with decompensated cirrhosis had a higher risk of ICU mortality compared to patients with compensated cirrhosis, independently of CHC status. During cirrhosis, sepsis is accompanied by a markedly imbalanced cytokine response with increased tissue damage and inflammation [14]. Moreover, patients with end-stage liver disease have enhanced intestinal permeability and translocation of bacteria and their products, imbalanced immune response, and aggravated intrahepatic microcirculatory dysfunction, which causes accumulation of toxins and immune dysfunction that may perpetuate end-stage organ dysfunction [34], and may explain the higher mortality rates [35].

As discussed above, HCV infection is a major cause of cirrhosis [36], and cirrhotic patients in ICUs have higher risks of sepsis and death, particularly in end-stage liver disease [15]. In our study, CHC appears to be a key factor that increases the risk of death in compensated cirrhotic patients with severe sepsis in ICUs, but the effect of CHC status was not maintained in other patients with compensated cirrhosis/non-severe sepsis or patients with decompensated cirrhosis (with or without severe sepsis). However, we have no clear explanation for this effect observed exclusively in patients with compensated cirrhosis and severe sepsis. On the one hand, CHC-associated comorbidities could promote negative clinical outcomes among cirrhotic patients, thus causing a significant proportion of the mortalities of CHC patients [37]. On the other hand, persistent HCV infection is the result of a series of connected events that culminate in diminished immunity and the inability to eliminate HCV infection [38]. This immune dysfunction is accompanied by inflammation and immune activation during CHC [38], which could be crucial in the prognosis of CHC-cirrhotic patients. Thus, it could be possible that CHC increases the immune dysfunction that occurs with severe sepsis, worsening the prognosis of cirrhotic patients. In this context, the effect of CHC could be diluted in the presence of decompensated cirrhosis in patients with a more advanced stage of disease, thus resulting in a worse prognosis, whereas the effect of CHC could be clearly observed in patients with compensated cirrhosis, as observed in our study.

Several points should be taken into account for the correct interpretation of our results (Limitations of the study):

Firstly, this study was retrospective and the acquisition of some clinical data related to HCV infection (viral status, patients who received prior treatment by antivirals, etc.) and the ICU (community acquired or nosocomial nature of the sepsis, or prognostic scores such as Child-Pugh, MELD, SOFA or CLIF-SOFA score) were unavailable from the MBDS records. Furthermore, we do not know the reason for admission of these patients; therefore, the prognostic could have been different if the patients were admitted for gastrointestinal bleeding, or for severe sepsis or septic shock. Neither do we know if these patients were admitted due to severe sepsis or acquired the sepsis in the ICU prior to death.

Secondly, the time until death or discharge was calculated from entry into the hospital rather than the ICU because the date of ICU admission was not recorded in the MBDS. A time bias could have occurred [37], since some patients may have been admitted directly into an ICU whereas others may have had a period of time between hospital admission and entry into an ICU, and this time period was included in the observation time. However, this theoretical bias should be well-balanced between the groups, and should only affect survival time.

Thirdly, due to the use of the administrative databases, was the inaccuracy in differentiating the etiologies of the diseases and the reporting of organ dysfunction, which could have engendered a confusion bias. In this context, grouping of ICD-9-CM codes into comorbidities, organ dysfunction, and site of infection (Additional file 1: SDC-Appendixes 1–6) may have been the best approach to solve this issue. On the other hand, MBDS has already proven its usefulness in the previous assessments of outcomes among patients admitted to ICU [39, 40]. This database has advantages being a national clinical administrative database, which allows analyze the trends in important public health issues.