This study demonstrated a high mortality rate among inpatients with liver cirrhosis and concomitant COVID-19 infection – 51.3%, compared to 21.7% in the control group, which did not have liver disease. We also observed a higher frequency of other adverse outcomes in patients with cirrhosis, including intensive care admission, renal replacement therapy, nosocomial infection, and VTE, as well as a longer length of stay when compared to matched controls. Furthermore, we found that patients with cirrhosis were more often referred for palliative care.
These findings align with previous research demonstrating a higher risk of mortality among patients with cirrhosis and COVID-19, as reported by Singh et al. in a cohort study conducted in the United States, which reported a relative risk (RR) of death 4.6 times higher for patients with cirrhosis [22]. In our study, we found that more than 50% of patients with cirrhosis have died, with 2.36-fold higher RR compared to those without cirrhosis. Ioannou et al. demonstrated in a large cohort involving veteran affairs that patients with cirrhosis and COVID-19 were more frequently hospitalized and had higher mortality than patients without cirrhosis [23]. However, there were no differences between both groups (cirrhosis and no cirrhosis) concerning mortality in the hospitalized subgroup analysis. Other studies reported higher 30-day mortality rates in cirrhotic patients [7, 23–26], as well as higher mortality rates considering inpatient plus hospice deaths compared with patients with COVID-19 without cirrhosis [27]. Most of these included patients in the pandemic's pre-vaccination phase. Our findings extend these results, showing that even after vaccination efforts, patients with cirrhosis still have a worse prognosis.
A large Swedish cohort study, which performed a propensity score-matching analysis, did not find a difference in the development of severe COVID-19 concerning mortality when comparing patients with and without cirrhosis [28]. Another cohort study, involving multiple European countries and using a large multinational database of patients with COVID-19, also conducted a propensity score-matched design [29]. This study included patients who had COVID-19 between March 2020 and March 2021 and found no difference in mortality after matching the cirrhosis group to the control group. The latter study included 70 patients with cirrhosis. These findings differ from our study, which also used propensity score-matched analysis to reduce bias due to other comorbidities, infection year, and other variables. Our findings are compatible with those from a large multinational cohort in the United Kingdom, which showed higher mortality in patients with cirrhosis and SARS-CoV-2 infection when compared to non-cirrhotic patients − 32% versus 8% [5]. However, after stratifying the cirrhosis group according to Child-Pugh stage-based classification and applying propensity score-matched models for each group, higher mortality rates were found in comparisons between Child B and Child C versus non-cirrhotic groups, while no difference was observed between Child A and the control group.
Despite those controversial findings involving patients with cirrhosis and COVID-19, some recent studies have also shown that patients at different stages of liver disease may exhibit different responses. Mallet et al in a large French cohort found that patients with decompensated liver cirrhosis had a significantly higher mortality rate, whereas patients with mild liver disease or compensated cirrhosis were not at increased risk of COVID-19-related death [8].
Our study adds to this evidence by highlighting a high incidence of other severe outcomes, including ICU admission, septic shock, dialysis, respiratory failure, and IMV. In addition, we observed that respiratory failure occurred more frequently in patients with advanced CLD in line with previous data, which suggests that it remains the leading cause of death among that group of patients [5].
Some factors could contribute to higher mortality in the patients with cirrhosis. One of them is the increased incidence of CKD in the case group since it is well-known that renal impairment in patients with COVID-19 can lead to worse outcomes [30, 31]. Furthermore, the higher incidence of smoking in the study group compared to the control group might also have contributed to the increased mortality, as smoking raises the risk of pulmonary lesions. Therefore, these patients may exhibit an exacerbated pulmonary response to COVID-19 infection.
In our study, a higher number of patients with cirrhosis were eligible for palliative care than in the control group. This observation likely reflects the severity of ACLF in these patients. As a large number of cirrhotic patients often necessitated intensive care interventions, we suppose that there was a delay in the transition to palliative care and the limitation of life-sustaining treatments. This delay in palliative care referral could have contributed to the over intervention observed in the cirrhosis group, underscoring the importance of timely and proactive palliative care management in this vulnerable population during acute illness episodes.
Cirrhosis is a prevalent disease in Brazil, especially alcohol-related hepatic cirrhosis [32]. Studies involving COVID-19 patients with underlying cirrhosis are scarce in Latin America. Thus, data from this study can help clinical professionals to better understand the characteristics of COVID-19 in this population.
However, this study has some limitations. Our statistical analysis did not account for the clinical behavior given the different strains prevalent at various phases of the pandemic (including the impact of vaccination). Additionally, the stage of liver disease was not considered in the analysis due to a lack of registered data on this basis. Moreover, this is a retrospective observational cohort study, which inherently carries limitations in reviewing patient records. Nonetheless, periodic audits were conducted to ensure data quality.
On the other hand, as a study strength, the utilization of advanced propensity score matching techniques enhances the robustness of our findings by minimizing selection bias and confounding effects, thus providing a more accurate estimation of the impact of cirrhosis on COVID-19 outcomes. This methodological rigor underscores the validity and reliability of our study's conclusions, contributing valuable insights to the existing body of knowledge. Furthermore, our study features an expressive number of patients across 41 different hospitals. The geographical diversity of hospitals across various regions of Brazil guarantees a diverse representation of the population in the study. Additionally, we analyzed data from the pre and post-vaccination phases. Most studies involving patients with cirrhosis were only developed in the first phase of the pandemic, with non-vaccinated patients.