This registry is the only French multicenter “real life” cohort, mainly prospective, and the largest one, describing 1219 SARS-CoV-2-infected patients with underlying chronic liver disease. Indeed, other international registries have been reported, the largest of which, reported by Marjot et al., included 745 patients followed for chronic liver disease but excluded liver transplant patients (3). Other concordant results come from public health databases (6, 14, 15, 16, 17) including a French retrospective study (6) based on Hospital Discharge database without detailed clinical characteristics of the patients.
The primary outcome confirms the major prognostic impact of advanced liver fibrosis on COVID-19-related short-term mortality. In particular, the results confirm the excess risk of death induced by the presence of cirrhosis, a fortiori by liver failure (assessed by the Child‒Pugh score B or C). Thus, advanced fibrosis was the main independent prognostic factor in this observational study, with an odds ratio of 2.688, regardless of the Child‒Pugh score. Age was also a hazardous factor in the entire population and the subgroups studied.
These results confirm previous data from other registries and cohorts that also showed an excess mortality of patients with cirrhosis of approximately 30% (3, 14, 15). In the French national hospital discharge database report by Mallet et al. (6), it seemed that the reduced access to intensive care and mechanical ventilation might have resulted in the excess mortality of patients with cirrhosis.
Nevertheless, according to Brozat et al. (14), the excess mortality of cirrhotic patients could be explained in part by the frequent comorbidities in these patients (diabetes, obesity, hypertension, chronic renal failure, etc.). On the other hand, a Swedish study of a population of patients with chronic liver disease showed that the presence of chronic liver disease was associated with an increased risk of hospitalization for COVID-19, but COVID-19-related mortality was not increased (16).
This result was a key argument behind our French liver society convincing the national health authorities to prioritize vaccination of patients with compensated cirrhosis in March 2021 before vaccination became universal.
The other major finding of the study was the absence of excess mortality risk in the subpopulation of immunocompromised patients, particularly in the subgroup of liver transplant patients, which represented 22% of the overall population, probably because, after transplant, the recipient has a normal and functional liver. This study is the largest cohort of liver transplant recipients with COVID-19 and confirms the results of several studies on liver transplant patients (10, 11, 18), in contrast to the results on kidney transplant patients, which have been associated with an increase in mortality related to COVID-19, probably because of their numerous associated comorbidities (chronic kidney failure, cardiovascular, type 2 diabetes, arterial hypertension, advanced age) and their high level of immunosuppression. There was no excess mortality in patients followed for autoimmune hepatitis, confirming the data of Marjot et al. (19).
Regarding the etiology of chronic liver disease, we found alcohol to be a factor associated with excess mortality in multivariate analysis, as did the registry studies of Marjot et al. and the French database from Mallet et al. (3, 6).
Other studies have shown that the prevalence of MAFLD (metabolism-associated fatty liver disease) ranges from 28 to 37% of patients infected with SARS-CoV-2 (14). MAFLD was associated with greater severity of COVID-19 (20), particularly in the subgroup of patients under 65 years of age in the meta-analysis by Wang et al. (13), and with increased mortality in these patients, with an OR of 2.93 (95% CI 1.87–4.6) (21), though there were no details on the degree of liver fibrosis.
The presence of a primary liver tumor was a collected data, but was not associated with mortality in our study. Some studies have investigated the impact of COVID-19 on the management and prognosis of HCC, one large international study showing a change in management (screening, diagnosis or treatment) in 80% of cases (22). In a French study of 670 patients (in two periods, 2019 vs. 2020), the number of new HCCs presented to the multidisciplinary consultation meeting was lower during the COVID-19 period, and the time to treatment was extended by one month in 20% of patients (23). More generally, the number of patients newly treated for digestive cancer, especially for HCC, fell drastically during the lockdown period (-42%) in a large population of patients over 65 years of age (24).
Our study has several limitations. The registry is not exhaustive in part because the inclusions were nonconsecutive and made by voluntary investigators, thus leading to a selection bias. Moreover, the observational was partially retrospective and was of a declarative nature, resulting in a large number of missing data, particularly biological data, which must be taken into account in interpreting the results.
Thus, although data from several large studies, including a meta-analysis of 64 studies involving more than 11,000 patients, showed that abnormalities on liver tests, mainly cytolysis, were frequent (> 20% of infected patients) (9) and were associated with excess mortality in hospitalized infected patients (8), in our study the biochemical data and the prognostic effect of disturbances of liver homeostasis were not interpretable.
Moreover, fibrosis was not assessed by the same methods in patients and only 26% of patients had liver biopsy as a reflect of nowadays clinical practice.
We did not collect information regarding the occurrence of post-COVID-19 cholangitis in the follow-up of patients, as there were too many missing biological data. This complication of chronic post-COVID-19 cholangitis has been regularly reported in the literature (25, 26, 27). The pathophysiology could be multifactorial, combining at least a direct viral toxicity, ischemic cholangitis similar to resuscitation cholangiopathies and a drug toxicity, in particular from ketamine.
Finally, as the observation was initiated before the start of universal vaccination programs against SARS-CoV-2, information on the vaccination status of the patients was not available. In addition, given the inclusion period, a minority of patients had an optimal vaccination schedule as currently recommended.