Liver injury in critically ill patients with COVID-19: a case series

Dear Editor,

Almost all reports on liver injury in patients with 2019 coronavirus disease (COVID-19) found blood liver tests to be frequently abnormal, especially in patients with more severe disease, but with substantial heterogeneity [1]. Moreover, blood liver tests’ abnormalities were frequently thought to be of doubtful clinical value.

Most studies have described blood liver tests in a single time point, usually at inclusion [2, 3]. Therefore, we used our case series of the first 20 consecutive patients with COVID-19 admitted to the intensive care unit (ICU) at Curry Cabral Hospital in Lisbon, Portugal, from March 10, 2020, onwards, to describe the temporal evolution of blood liver tests.

Median (interquartile range (IQR)) age was 67 (52–74) years with 18 (90%) males (Table 1). Median (IQR) time from symptom onset to hospital admission was 7.5 (5.5–8.5) days, and median time from hospital admission to ICU admission was 1.1 (0.7–2.1) days. All patients required invasive mechanical ventilation on ICU admission. Median (IQR) sequential organ failure assessment (SOFA) score on ICU admission and peak was 8 (7–9) and 9 (8–11), respectively. As of April 10, following a median (IQR) of 21.5 (11.2–25.4) days post ICU admission, 3 (15%) patients died of multi-organ failure, 14 (70%) were discharged to the ward, and 3 (15%) remained in the ICU.

Table 1 Baseline characteristics and outcomes of patients

No patient had documented liver disease prior to hospitalization. During the first 10 days post ICU admission, all patients had at least one abnormal blood liver test. Overtime, only median gamma-glutamiltranspeptidase (GGT), alanine transferase (ALT), and aspartate transferase (AST) showed any increase from upper limit of normal (ULN) and only median GGT had a ≥ 3 fold increase from ULN (Table 2). Median peak GGT was on day 8 post ICU admission. Patients with peak C-reactive protein ≥ 250 mg/L (day 4 post ICU admission) had higher but non-significant median peak GGT (298 vs. 125 IU/L), ALT (101 vs. 42 IU/L), or AST (72 vs. 57 IU/L) than others (P > 0.50 for all comparisons).

Table 2 Temporal evolution of median levels of blood tests

In our case series, liver injury was frequent but generally transient and non-severe. While synthetic function was largely preserved, late cholestasis was frequently observed. Cholestasis may have been associated to the critical illness itself (inflammation), parenteral nutrition (only 2 patients required parenteral nutrition), or drug toxicity (all patients were on antibiotics, for example ceftriaxone, amoxicillin-clavulanate, or azithromycin) [4, 5]. While cholestasis could have been multifactorial, the differential diagnosis was not easy to perform with precision. Overall, attention to modifiable factors, such as control of inflammation, timing of parenteral nutrition, and avoiding drugs with worse liver toxicity profile, may be important to prevent cholestasis progression in these patients. Further studies are needed to understand liver injury in critically ill patients with COVID-19, especially if there is any direct viral effect on the liver cells.