Letter: liver disease and COVID-19-not the perfect storm.

EDITORS,

We read with great interest the article by Garrido et al on COVID‐19 and liver disease. We agree that COVID‐19 is associated with elevated liver enzymes, but its significance needs to be further elucidated as well as the impact of underlying liver disease on COVID‐19. From 535 patients SARS‐CoV‐2 patients admitted at Imperial College Healthcare NHS Trust, London, from 25 February to 5 of April 2020, we found 27.7% (148/535) had elevated liver transaminases. Median ALT and AST levels were 48 IU/L (IQR 17‐49) and 88 IU/L (52‐172) respectively. Only 1.7% (9/535) had an ALT five times the upper limit of normal and no patients developed synthetic liver dysfunction nor acute liver failure suggesting COVID‐19 is not associated with significant acute liver injury. Patients with elevated liver transaminases had similar mortality rates than patients with normal transaminases at admission (29.4% vs 23.3%; P = 0.462), although a greater proportion with abnormal liver transaminases were admitted to intensive care (29.8% vs 10.7%; P < 0.001), in line with previous findings. We also did not find a correlation between the degree of transaminitis and hypoxaemia on admission defined by a blood oxygen saturation level < 94% (r = 0.06, P = 0.42) suggesting no direct link between COVID‐19‐induced initial hypoxemia and liver damage.

In addition, we have explored the impact of pre‐existing liver disease—including cirrhosis, on the outcomes of COVID‐19, which we feel is of great importance. Immune dysfunction has been well described in cirrhotic patients resulting in a high risk of infection and cirrhotic patients have poor outcomes with acute respiratory distress syndrome (ARDS). In our cohort, 89/535 (16.6%) had suspected pre‐existing liver disease based on radiological and laboratory results, including 21/535 (3.9%) with cirrhosis, supporting previously reported rates, and 13/21 (61.9%) had decompensated cirrhosis (Child‐Pugh B/C) (Table 1). The main aetiology of cirrhosis was alcohol (47.6%). COVID‐19 symptoms at admission, in cirrhotic patients, were mainly cough (71.4%), fever (57.1%) and shortness of breath (23.8%), rather than worsening features of hepatic decompensation. Following admission, 6/21 (28.6%) with compensated cirrhosis developed hepatic decompensation (ie new onset encephalopathy, ascites, variceal haemorrhage or jaundice). This hepatic decompensation rate was not higher than that observed in historical in‐hospital pre‐COVID cirrhosis data (28.6% vs. 37%).

TABLE 1

Baseline characteristics of the study population according to the pre‐existing liver disease. Data are median (IQR) or n (%). Statistical test used for categorical data assessed using Chi‐squared test. Statistical test used for continuous, nonparametric variable assessed using Kruskal‐Wallis test

Variable	Patients with cirrhosis (n = 21)	Patients with pre‐existing noncirrhotic liver disease (n = 68)	Patients without pre‐existing liver disease (n = 446)	P
Age – years	71 (57‐83)	61 (53‐76)	67 (54‐79)	0.115
Gender – no (%), Male sex	14/21 (67.0%)	40/68 (59.0%)	276/446 (61.9%)	0.976
Ethnicity – no (%)
Asian	4/21 (19.0%)	13/68 (19.1%)	51/446 (11.4%)	0.021
Black	2/21 (9.5%)	16/68 (23.5%)	77/446 (17.3%)
White	10/21 (47.6%)	25/68 (36.8%)	149/446 (33.4%)
Other	5/21 (23.8%)	14/68 (20.6%)	169/446 (37.9%)
Observations on admission
BMI – kg/m2 *	28.1 (21.8‐31.3)	30.3 (26.2‐34.0)	27.4 (23.6‐31.8)	0.023
Systolic Blood Pressure (SBP) – mm Hg	125 (121‐154)	129 (114‐148)	131 (114‐152)	0.869
Diastolic Blood Pressure (DBP) – mm Hg	73 (65‐84)	75 (69‐87)	75 (65‐83)	0.750
Heart Rate – beats per minute	82 (74‐93)	92 (68‐151)	91 (80‐104)	0.026
Respiratory Rate per minute	20 (16‐24)	20 (18‐28)	20 (18‐26)	0.302
Temperature – °C	36.6 (36.0‐37.7)	37.5 (36.6‐38.5)	37.2 (36.6‐37.9)	0.108
Overall in‐hospital mortality – no (%)	8/21 (38.1%)	22/68 (32.4%)	151/446 (34.1%)	0.689
Haemoglobin, g/L	116 (103‐150)	130 (112‐144)	134 (119‐145)	0.008
Platelet count, 109/L	113 (74‐209)	205 (155‐282)	199 (157‐261)	<0.001
Lymphocyte count, 109/L	0.9 (0.6‐1.2)	1.0 (0.7‐1.5)	0.9 (0.6‐1.3)	0.287
Creatinine, µmol/L	94 (60‐174)	90 (69‐127)	91 (72‐126)	0.944
Urea, mmol/L	6.6 (4.0‐18.2)	6.3 (4.2‐11.0)	6.2 (4.1‐10.4)	0.757
Total bilirubin, µmol/L	31 (12‐63)	9 (8‐16)	11 (8‐15)	<0.001
ALT, IU/L	32 (14‐50)	30 (16‐57)	28 (17‐49)	0.969
AST, IU/L	127 (83‐170)	71 (45‐183)	106 (71‐193)	0.865
ALP, IU/L	152 (101‐259)	92 (61‐127)	79 (63‐103)	<0.001
Albumin, g/L	25 (20‐30)	31 (25‐35)	31 (27‐34)	0.002
Ferritin, µg/L	437 (158‐1104)	889 (440‐1916)	796 (430‐1762)	0.092
CRP, mg/L	53.1 (27.8‐161.1)	79.1 (27.0‐183.5)	109.9 (52.6‐189.2)	0.037
d‐dimer, ng/ml	2142 (1475‐2484)	1308 (764‐2089)	1351 (714‐2592)	0.476
Prothrombin time, s	14.1 (13.1‐25.1)	14.3 (13.2‐15.8)	13.8 (13.1‐14.8)	0.262
Troponin, ng/L	13 (9.5‐32.5)	16.5 (5.0‐45.8)	13 (5‐48)	0.967
BNP, ng/L	149 (58‐421)	23 (10‐64)	38 (13‐114)	0.047

The overall mortality of COVID‐19 patients with cirrhosis was similar to those COVID‐19 patients without liver disease (38.1% vs. 34.1%, P = 0.689). Median duration of admission was longer in cirrhotic patients compared to those without pre‐existing liver disease (10 days vs. 7 days, P = 0.016), but development of ARDS requiring admission to ICU for intubation and ventilation was not statistically significant (5.3% vs. 13.7%, P = 0.280). At 14 days after diagnosis, hospital readmission rates for those with cirrhosis were comparable with those without liver disease (14.3% vs 11.8%, P = 0.78) and only one patient re‐presented to hospital with bacterial superinfection.

To summarise, our findings suggest that: i. one third of COVID‐19 patients have mild transaminitis, which may be associated with more severe disease but not increased mortality; ii. liver cirrhosis does not predispose to increase mortality but does increase length of stay. Further outcome data from large registries of COVID‐19 patients with cirrhosis are awaited to confirm this.