On-admission LBAI associated with cumulative ALI
Liver function on admission predicted the subsequent risk of cumulative occurrence of ALI among COVID-19 patients to a great extent, as shown in figure 1. The cumulative incidence of ALI over two months was 41.9% (95% CI 31.4 to 51.9%) in patients with on-admission LBAI, dramatically higher than the 9.6% (95% CI 4.7 to 16.5%) in patients with on-admission normal LB (p <0.0001); the risk was 8-fold higher in the LBAI group than that in the normal LB group (HR 8.07; 95% CI 4.23 to 5.37). We also observed similar patterns albeit with moderately reduced effect (HR 4.83; 95% CI 2.45 to 9.54) when on-admission LBA associated with the subsequent risk of in-hospital incidence of ALI was analyzed (supplementary figure 1).
Drugs and treatment associated with LBAI
To address the issue of whether liver dysfunction was caused by drugs commonly used to treat COVID-19, such as antibiotics, antivirus, traditional Chinese medicine/Chinese patent drug, and intravenous nutrition (IVN), we analyzed the association of drugs with subsequent incidence of ALI or LBAI or worsening LB (figure 2 and supplementary figure 2). The results showed that patients who received IVN either on admission or in hospital were associated with cumulative LBAI or ALI; a similar association was found between in-hospital use of antibiotics with subsequent ALI or worsening LB.
Treatment with oxygen therapy by mask, high-flow nasal cannula, or mechanical ventilation were used approximately 6-fold more in patients with cumulative LBAI than those with normal cumulative LB, indicating hypoxia was an important risk factor of LBAI (supplementary table 1).