Chronic liver disease independently associated with COVID-19 severity: evidence based on adjusted effect estimates

Abstract

Recently, a meta-analysis by Kovalic et al. reported that chronic liver disease was significantly associated with severe coronavirus disease 2019 (COVID-19) and mortality [1]. This is an interesting study. However, the pooled effect on the association between chronic liver disease and severe COVID-19 was estimated based on un-adjusted effect sizes in Kovalic et al.’s study [1]. It has been considered that several factors such as gender, age and certain comorbidities significantly influenced COVID-19 outcomes [2–5]. This suggests that these factors might modulate the relationship between chronic liver disease and COVID-19 severity. Therefore, it is urgently required to clarify this association by performing a quantitative meta-analysis based on adjusted effect estimates. Electronic databases including PubMed, Web of Science and EMBASE were searched up to December 10, 2020 using the terms: “SARS-CoV-2”, “COVID-19”, “chronic liver disease”, “cirrhosis”, “hepatitis”, “liver cancer” and “nonalcoholic fatty liver disease”. Only studies reporting the relationship between chronic liver disease and COVID19 severity by adjusted effect estimates were included. Case reports, reviews, duplicate publications, errata and studies without sufficient data were excluded. The heterogeneity was detected by I2 statistics. The pooled effect sizes with 95% confidence interval (CI) were estimated. Publication bias was evaluated by Begg’s test and Egger’s test. Sensitivity analysis, subgroup analysis and meta-regression analysis were also performed. All data were analyzed using Stata 12.1. p < 0.05 was considered statistical significance. Figure S1 shows the flow diagram of study selection. 29 articles with 90,095 confirmed COVID-19 patients were included. The characteristics of the included studies are summarized in Table 1. Our meta-analysis based on adjusted effect estimates demonstrated that COVID-19 patients with chronic liver disease tended to develop severe outcome compared to those without (pooled effect size = 1.52, 95% CI: 1.14–2.02, Fig. 1a) and had a significantly increased risk for mortality compared to those without (pooled effect size = 1.36, 95% CI: 1.22–1.53, Fig. 1b). Sensitivity analysis exhibited that our findings were stable (Fig. 1c). Subgroup analyses by sample size and study design exhibited consistent results (Table S1 and Figure S2–3). But inconsistent results were observed in subgroup analyses by age, male percentage, effect estimate and region (Table S1 and Figure S4–7). Meta-regression analysis showed that the tested variables such as sample size, age, male percentage, effect estimate, study design and region might not be the source of heterogeneity (Table S1). Begg’s test and Egger’s test suggested that there might be potential publication bias (Figure S8). This meta-analysis has several limitations. First, inconsistent results were observed in subgroup analyses by age, male percentage and region. Thus, the findings should be cautiously extrapolated to whole population. Second, most of the included studies are retrospective, further welldesigned studies with more prospective literatures are warranted to confirm our findings. Third, publication bias might Supplementary Information The online version contains supplementary material available at https ://doi.org/10.1007/s1207 2-020-10133 -y.