Liver injury associated with the severity of COVID-19: A meta-analysis

Ruiqi Yang, Jihua Feng, Huan Wan, Xiaona Zeng, Pan Ji, Jianfeng ZhangFrontiers in Public Health 2023 · DOI

Abstract: https://www.crd.york.ac.uk/prospero/, Identifier: CRD42022325206.

Introduction

Since the emergence of the COVID-19 pandemic, clinical observations have frequently highlighted abnormal liver enzyme levels in infected patients. However, the exact nature and clinical significance of these abnormalities remained a subject of debate. Researchers questioned whether liver damage was a direct viral effect, a consequence of aggressive pharmacotherapy, or a secondary result of systemic inflammation and hypoxia.

This meta-analysis was conducted to provide a definitive, large-scale evaluation of the association between liver injury and COVID-19 severity. By synthesizing data from dozens of global studies, the authors aimed to determine if liver chemistry could serve as a reliable "early warning system" for identifying patients at high risk of progressing to severe disease or death.

Methodology

The researchers performed a systematic review and meta-analysis following PRISMA guidelines. They searched major medical databases, including PubMed, Embase, and the Cochrane Library, for studies published through August 2022.

Key technical aspects of the approach included:

  • Statistical Modeling: A random-effects model was utilized to calculate Pooled Odds Ratios (OR) for categorical outcomes (such as mortality) and Weighted Mean Differences (WMD) for continuous laboratory values. This model accounts for the high degree of heterogeneity expected between diverse clinical studies.
  • Quality Assessment: The quality of included studies was rigorously vetted using the Newcastle-Ottawa Scale (NOS), ensuring that only high-quality data influenced the final results.
  • Bias Detection: Egger’s test and funnel plots were employed to check for publication bias, ensuring the results were not skewed by the selective publication of "positive" findings.

Datasets and Study Selection

The study synthesized data from 46 individual studies involving a total of 28,663 patients. The cohort was comprised of 6,310 severe cases and 22,353 non-severe cases.

PRISMA flow diagram illustrating the systematic selection process of the 46 included studies.

While the majority of studies (33) originated in China, the meta-analysis also included data from Japan, Austria, the USA, France, Spain, Iran, Ethiopia, and Brazil, providing a broad global perspective on the phenomenon.

Key Insights

  • Prognostic Value of Liver Markers: Liver function markers (ALT, AST, GGT, TBIL, and Albumin) appear highly sensitive to the systemic "storm" caused by COVID-19. Monitoring these levels early can assist clinicians in triaging patients before respiratory failure becomes critical.
  • Bile Duct Vulnerability: SARS-CoV-2 enters human cells via the ACE2 receptor. While hepatocytes have low ACE2 expression, bile duct epithelial cells express it highly. This explains why markers like GGT and Total Bilirubin were significantly elevated in severe cases, suggesting the virus may directly target the biliary system.
  • Iatrogenic Factors: A strong correlation was found between liver injury and the use of antiviral drugs (e.g., lopinavir/ritonavir) and antibiotics. This suggests that some observed liver damage may be drug-induced, emphasizing the need for careful medication management.
  • Synthetic Function Impairment: Severe cases exhibited significantly lower Albumin levels and prolonged Prothrombin Time (PT). This indicates that the liver’s ability to synthesize essential proteins and clotting factors is compromised, contributing to multi-organ failure.

Key Results

The quantitative findings demonstrate a clear link between liver involvement and COVID-19 severity:

  • Mortality Risk: Patients with liver injury faced a 2.72 times higher risk of death (OR = 2.72, 95% CI: 1.18–6.27).
  • Prevalence of Injury: The odds of having liver injury were 3.25 times higher in severe COVID-19 patients compared to mild cases.
Forest plots showing the association between liver injury and COVID-19 severity (A) and mortality risk (C).
  • Enzyme Elevations: Severe cases showed significantly higher levels of key enzymes, including ALT (+12.87 U/L), AST (+13.98 U/L), and GGT (+20.67 U/L).
  • Metabolic Markers: Severe cases had higher Total Bilirubin (+2.98 µmol/L) and significantly lower Albumin (-4.52 g/L) compared to non-severe cases.

Discussion and Limitations

Despite the strong associations found, the authors acknowledge several limitations:

  • High Heterogeneity: Many metrics showed an $I^2$ value over 90%, indicating significant variation between studies likely due to differences in hospital protocols and patient demographics.
  • Retrospective Nature: Most included studies were retrospective, making it difficult to establish a strict cause-and-effect relationship.
  • Baseline Data Gaps: Many studies lacked information on patients' pre-existing liver health, complicating the distinction between new injury and the exacerbation of chronic conditions.

Conclusion: This meta-analysis confirms that COVID-19 significantly impairs liver function and that liver injury is a powerful prognostic indicator. Future research should focus on prospective, multi-center studies that adjust for age, sex, and specific drug treatments to further isolate the direct impact of the virus on liver tissue.

Summary updated

Abstract

https://www.crd.york.ac.uk/prospero/, Identifier: CRD42022325206.
BibTeX
@inproceedings{yang2023liver,
  author = {Ruiqi Yang and Jihua Feng and Huan Wan and Xiaona Zeng and Pan Ji and Jianfeng Zhang},
  title = {{Liver injury associated with the severity of COVID-19: A meta-analysis}},
  booktitle = {Frontiers in Public Health},
  year = {2023},
  doi = {10.3389/fpubh.2023.1003352},
}
Figure