Introduction
In December 2019, the outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV 2) virus spread globally to the extent of becoming a pandemic. In Malaysia, the first confirmed case of coronavirus disease of 2019 (COVID-19) was detected on 23 February 2020 and until now there have been 3.5 million cases nationwide, with a lower mortality rate of 0.91%.1 Over 2 years, multiple variants of the original virus have emerged. Variants of concern (VoCs) include Alpha, Beta, Gamma, Delta, and Omicron.2 The Delta variant is recognized as more infective than the other VoCs and causes more severe infections and more deaths.3–5 The first case of Delta variant infection was detected in Malaysia on 19 April 2021 and continued to be the predominant strain until January 2022.1
SARS-CoV-2 is known to primarily cause upper respiratory tract infections that lead to lower respiratory tract infections in a subset of patients, causing alveolar damage and subsequent respiratory failure.6–8 Although respiratory tract involvement is the prominent clinical manifestation, liver dysfunction in COVID-19 patients is a recognized entity, and the cause of this phenomenon is multifactorial.9–15 The incidence of abnormal liver biochemistry (ALB) is common (14–76.3%) and varies widely among populations.11–14,16–18 These studies also revealed that disease severity, increasing age, and underlying chronic liver disease (CLD) were factors that increased the risk of liver injury (LI). However, data on the extent of liver dysfunction caused by different VoCs are lacking. There is also a paucity of local data regarding factors that increase the risk of liver dysfunction in people with COVID-19 infection. This study aimed to determine (1) the prevalence and risk factors of ALB and LI in adult COVID-19 patients and (2) the differences of ALB and LI in patients with the Delta variant compared with the wild-type and all VoCs that appeared prior to the Delta variant in Malaysia.
Materials and methods
Study design and participants
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki and was approved by the Research Ethics Committee of University of Malaya (MECID No. 202146-10036). All subjects gave written informed consent. We performed a prospective observational study and recruited patients who were admitted between 1 February 2020 and 30 October 2021 using a convenience sampling method. Patients 18 years of age and older who were admitted to one hospital in Malaysia with laboratory-confirmed COVID-19 confirmed by real-time reverse-transcriptase polymerase chain reaction (RT-PCR) assay of nasal and pharyngeal swab specimens and had at least a single liver function test (LFT) during hospitalization were recruited into the study. Patients were divided into cohort A (admission between February 2020 and 30 April 2021), which included the COVID-19 wild-type strain and all VoCs before the Delta variant and cohort B (admission between 1 May 2021 and 31 October 2021), which included COVID-19 Delta variant cases. Owing to lack of genomic sequencing, COVID-19 cases detected from May 2021 onwards were presumed to be of the Delta variant, as the first case of Delta variant infection in Malaysia was detected on 12 April 2021 and cases increased rapidly from May 2021 onward.
Data collection
Relevant clinical data, laboratory, and imaging results were obtained by reviewing the electronic medical record of the patients. Information regarding the presence of liver disease diagnosed either before or during admission, pharmacotherapy, disease severity, and outcome were documented. CLD was defined when a patient had a diagnosis of liver cirrhosis, chronic hepatitis B, or C infection, alcoholic liver disease, nonalcoholic steatohepatitis, and/or autoimmune liver disease.
ALB was defined and categorized into hepatocellular type when alanine aminotransaminase (ALT) or aspartate aminotransferase (AST) was more than 40 U/L, cholestatic type when alkaline phosphatase (ALP) was more than 130 U/L or gamma-glutamyl transferase (GGT) was more than 50 U/L, or mixed type if both patterns were present. LI was defined and categorized as hepatocellular if AST or ALT was more than three times the upper limit of normal, cholestatic type if ALP or GGT or total bilirubin was more than two times the upper limit of normal, or mixed if both patterns were present. For patients with more than one LFT result during hospitalization, the most deranged LFT result was documented. COVID-19 disease severity was categorized as mild (category 1–2), moderate (category 3), or severe (category 4–5). Definitions of the categories were: category 1, asymptomatic COVID-19 patients; category 2, symptomatic patients with no evidence of pneumonia and not requiring oxygen support; category 3, patients with clinical or radiological evidence of pneumonia but not requiring oxygen support; category 4, patients with pneumonia and requiring oxygen support; and category 5, patients requiring mechanical ventilation or with organ failure.
Statistical analysis
Categorical variables were reported as numbers and percentages. Chi-square and Fisher’s exact tests were used to compare categorical variables. Continuous values were reported as means (standard deviations) and were compared with Student’s t-test or one-way analysis of variance for parametric data. Continuous values were reported as medians and were compared with Mann-Whitney U or Kruskal-Wallis H tests for nonparametric data. Multivariate analysis was performed using multiple linear regression. The statistical analyses were performed with SPSS version 21 (IBM Corp., Armonk, NY, USA). A two-sided significance level of p ≤ 0.05 was used for all models.
Results
We included 1,246 patients, 697 (58.7%) in cohort A and 549 (44.1%) in cohort B who were admitted to our hospital with COVID-19 infection between 1 February 2020 and 30 October 2021. The median age was 50.0 years with a male:female ratio of 1:1.02. More than half (52.1%) had mild disease and 47.1% were given specific treatment. CLD was present in 11.7% of patients. The baseline and clinical characteristics of these patients are outlined in Table 1.
Table 1Demographics and clinical characteristics of COVID-19 patients (n = 1,246)
| Characteristic | Cohort*
| p-value |
|---|
| All, n = 1,246 | Cohort A, n = 697 | Cohort B, n = 549 |
|---|
| Age in years | | | | |
| Median [IQR] | 50.0 (31) | 44.0 (31) | 55.0 (27) | <0.001 |
| 1–60 | 846 (67.9) | 508 (72.9) | 338 (61.6) | <0.001 |
| >60 | 400 (32.1) | 189 (27.1) | 211 (38.4) | |
| Sex | | | | |
| Male | 618 (49.6) | 325 (46.6) | 256 (53.4) | 0.018 |
| Female | 628 (50.4) | 372 (53.4) | 293 (46.6) | |
| Ethnicity | | | | |
| Malay | 779 (62.5) | 427 (61.3) | 352 (64.1) | 0.123 |
| Chinese | 233 (18.7) | 126 (18.1) | 107 (19.5) | |
| Indian | 189 (15.2) | 112 (16.1) | 77 (14.0) | |
| Others | 45 (3.6) | 32 (4.6) | 13 (2.4) | |
| COVID-19 severity | | | | |
| Mild | 649 (52.1) | 450 (64.6) | 199 (36.2) | <0.001 |
| Moderate | 200 (16.1) | 108 (15.5) | 92 (16.8) | |
| Severe | 397 (31.9) | 139 (19.9) | 258 (47.0) | |
| Disease-modulating pharmacotherapy | | | | |
| Yes | 587 (47.1) | 231 (33.1) | 356 (64.8) | <0.001 |
| No | 659 (52.9) | 466 (66.9) | 193 (35.2) | |
| Type of treatment | | | | |
| NSAID | 81 (6.5) | 68 (12.4) | 13 (1.9) | <0.001 |
| Hydroxychloroquine | 25 (2.0) | 22 (3.2) | 3 (0.5) | 0.001 |
| Favipiravir | 175 (14.0) | 56 (8.0) | 119 (21.7) | <0.001 |
| Methylprednisolone | 121 (9.7) | 40 (5.7) | 81 (14.8) | <0.001 |
| Clexane | 273 (21.9) | 73 (10.5) | 200 (36.4) | <0.001 |
| Tocilizumab | 110 (8.8) | 48 (6.9) | 62 (11.3) | 0.006 |
| Dexamethasone | 455 (36.5) | 164 (23.5) | 291 (53.0) | <0.001 |
| Baricitinib | 5 (0.4) | 0 (0.0) | 5 (0.9) | 0.012 |
| Comorbidities | | | | |
| CLD | 146 (11.7) | 75 (10.8) | 71 (12.9) | 0.237 |
| NAFLD | 126 (10.1) | 67 (9.6) | 59 (10.7) | 0.510 |
| DM | 388 (31.1) | 179 (25.7) | 209 (38.1) | <0.001 |
| HTN | 441 (35.4) | 205 (29.4) | 236 (43.0) | <0.001 |
| CKD | 91 (7.3) | 44 (6.3) | 47 (8.6) | 0.130 |
| CCF | 74 (5.9) | 42 (6.0) | 32 (5.8) | 0.884 |
| Chronic lung disease | 86 (6.9) | 42 (6.0) | 44 (8.0) | 0.169 |
| Cancer | 51 (4.1) | 31 (4.4) | 20 (3.6) | 0.477 |
| ALB | | | | |
| Yes | 731 (58.7) | 338 (48.5) | 393 (71.6) | <0.001 |
| No | 515 (41.3) | 359 (51.5) | 156 (28.4) | |
| Type of ALB | | | | |
| Hepatocellular | 205 (16.5) | 116 (16.6) | 89 (16.2) | <0.001 |
| Cholestasis | 80 (6.4) | 46 (6.6) | 34 (6.2) | |
| Mixed | 446 (35.8) | 176 (25.3) | 270 (49.2) | |
| Deranged LFT | | | | |
| Total bilirubin | 153 (21.2) | 61 (18.3) | 92 (23.6) | 0.084 |
| ALP | 135 (18.8) | 59 (18.0) | 76 (19.4) | 0.620 |
| ALT | 475 (65.1) | 192 (57.0) | 283 (72.0) | <0.001 |
| AST | 583 (82.3) | 237 (75.0) | 346 (88.3) | <0.001 |
| GGT | 391 (54.4) | 162 (49.2) | 229 (58.7) | 0.011 |
| Liver injury | | | | |
| Yes | 332 (26.6) | 119 (17.1) | 213 (38.8) | <0.001 |
| No | 914 (73.4) | 578 (82.9) | 336 (61.2) | |
| Type of liver injury | | | | |
| Hepatocellular | 54 (4.3) | 27 (3.9) | 27 (4.9) | <0.001 |
| Cholestasis | 149 (12.0) | 54 (7.7) | 95 (17.3) | |
| Mixed | 123 (9.9) | 37 (5.3) | 86 (15.7) | |
| Abnormal LFT | | | | |
| Total bilirubin | 86 (26.2) | 30 (25.4) | 56 (26.7) | 0.806 |
| ALP | 100 (30.3) | 39 (33.1) | 61 (28.8) | 0.418 |
| ALT | 267 (80.4) | 86 (72.3) | 181 (85.0) | 0.005 |
| AST | 306 (94.4) | 102 (91.1) | 204 (96.2) | 0.054 |
| GGT | 291 (88.4) | 97 (82.9) | 194 (91.5) | 0.019 |
| Liver failure | | | | |
| Yes | 14 (1.1) | 7 (1.0) | 7 (1.3) | 0.653 |
| No | 1,232 (98.9) | 690 (99.0) | 542 (98.7) | |
| ICU admission | | | | |
| Yes | 172 (13.8) | 84 (12.1) | 88 (16.0) | 0.043 |
| No | 1,074 (86.2) | 613 (87.9) | 461 (84.0) | |
| Ventilator | | | | |
| Yes | 354 (28.4) | 123 (17.6) | 231 (42.1) | <0.001 |
| No | 892 (71.6) | 574 (82.4) | 318 (57.9) | |
| Death | | | | |
| Yes | 61 (4.9) | 28 (4.0) | 33 (6.0) | 0.105 |
| No | 1,185 (95.1) | 669 (96.0) | 516 (94.0) | |
Prevalence and risk factors for ALB and LI
ALB was seen in 731 (58.7%) patients, but only 26.6% experienced LI and 14 (1.1%) patients had liver failure. Mixed type and cholestasis were the predominant patterns of liver insult seen among patients with ALB and LI, respectively; specifically, 69.3% above 60 years and 72% of male patients had ALB, and incidence increased with disease severity, and 78% of patients who received any disease-modulating pharmacotherapy had ALB. Notably, the majority of patients with comorbidities who required intensive care unit (ICU) admission, or died were found to have ALB as well (Table 2). Multivariate analysis showed a significantly higher risk of developing ALB in men (OR = 2.93, 95% CI: 2.24, 3.83), patients with moderate (OR = 1.66, 95% CI: 1.10, 2.50) or severe (OR = 3.17, 95% CI: 1.53, 6.57) disease, CLD (OR = 19.31, 95% CI: 2.38, 156.53), or hypertension (HTN) (OR = 1.64, 95% CI: 1.16, 2.33). In total, 34.8% of patients older than 60 years of age and 43.8% who received disease-modulating pharmacotherapy were reported to have LI. The frequency of LI also increased with disease severity. The majority of patients who required mechanical ventilation, ICU admission, or died also had LI (Table 3). Multivariate analysis for LI showed significantly higher frequencies in men (OR = 1.79, 95% CI: 1.33, 2.41), in patients with moderate (OR = 3.03, 95% CI: 1.87, 4.92) or severe (OR = 4.56, 95% CI: 2.27, 9.18) disease, or CLD (OR = 2.93, 95% CI: 1.04, 8.28). The risk was lower in Indian patients (OR = 0.61, 95% CI: 0.39, 0.95) compared with other ethnic groups.
Table 2Multivariate analysis of COVID-19 patients with abnormal liver biochemistry (n = 731)
| Characteristics | Overall
| Cohort A
| Cohort B
|
|---|
Abnormal Liver Biochemistry
| AOR (95% CI) | p-value | Abnormal Liver Biochemistry
| AOR (95% CI) | p-value | Abnormal Liver Biochemistry
| AOR (95% CI) | p-value |
|---|
| Yes | No | Yes | No | Yes | No |
|---|
| Age in years | | | | | | | | | | | | |
| ≤60 | 454 (53.7) | 392 (46.3) | – | – | 219 (43.1) | 289 (56.9) | 1.08 (0.65, 1.80) | 0.770 | 235 (69.5) | 103 (30.5) | 1.85 (1.05, 3.28) | 0.34 |
| >60 | 277 (69.3) | 123 (30.8) | 0.73 (0.51, 1.05) | 0.092 | 119 (63.0) | 70 (37.0) | – | – | 158 (74.8) | 53 (35.1) | – | – |
| Sex | | | | | | | | | | | | |
| Female | 286 (45.5) | 342 (54.5) | – | – | 131 (35.2) | 241 (64.8) | – | – | 155 (60.5) | 101 (39.5) | – | – |
| Male | 445 (72.0) | 173 (28.0) | 2.93 (2.24, 3.83) | <0.001 | 207 (63.7) | 118 (36.3) | 3.28 (2.32, 4.66) | <0.001 | 238 (81.2) | 55 (18.8) | 2.79 (1.77, 4.42) | <0.001 |
| Ethnicity | | | | | | | | | | | | |
| Malay | 440 (56.5) | 339 (43.5) | – | – | 196 (45.9) | 231 (54.1) | – | – | 244 (69.3) | 108 (30.7) | – | – |
| Chinese | 157 (67.4) | 76 (32.6) | 0.98 (0.67, 1.43) | 0.908 | 71 (56.3) | 55 (43.7) | 0.87 (0.53, 1.43) | 0.576 | 86 (80.4) | 21 (19.6) | 1.13 (0.59, 2.16) | 0.674 |
| Indian | 105 (55.6) | 84 (44.4) | 0.71 (0.48, 1.05) | 0.085 | 53 (47.3) | 59 (52.7) | 0.70 (0.43, 1.14) | 0.151 | 52 (67.5) | 25 (32.5) | 0.79 (0.41, 1.54) | 0.713 |
| Others | 29 (64.4) | 16 (35.6) | 1.52 (0.75, 3.07) | 0.241 | 18 (56.3) | 14 (43.8) | 1.51 (0.68, 3.35) | 0.314 | 11 (84.6) | 2 (15.4) | 2.22 (0.38, 12.95) | 0.492 |
| COVID-19 severity | | | | | | | | | | | | |
| Mild | 264 (40.7) | 385 (59.3) | – | – | 171 (38.0) | 279 (62.0) | – | – | 93 (46.7) | 106 (53.3) | – | – |
| Moderate | 123 (61.5) | 77 (38.5) | 1.66 (1.10, 2.50) | 0.015 | 61 (56.5) | 47 (43.5) | 1.29 (0.74, 2.25) | 0.363 | 62 (67.4) | 30 (32.6) | 1.85 (0.96, 3.57) | 0.067 |
| Severe | 344 (86.6) | 53 (13.4) | 3.17 (1.53, 6.57) | 0.002 | 106 (76.3) | 33 (23.7) | 2.05 (0.68, 6.19) | 0.201 | 238 (92.2) | 20 (7.8) | 4.54 (1.52, 13.54) | 0.007 |
| Treatment | | | | | | | | | | | | |
| 458 (78.0) | 129 (22.0) | 1.07 (0.63, 1.83) | 0.796 | 162 (70.1) | 69 (29.9) | 2.36 (1.10, 5.03) | 0.027 | 296 (83.1) | 60 (16.9) | 0.52 (0.21, 1.26) | 0.147 |
| Types of treatment | | | | | | | | | | | | |
| NSAID | 64 (79.0) | 17 (21.0) | 1.18 (0.60, 2.33) | 0.633 | 176 (37.8) | 290 (62.2) | 0.55 (0.15, 2.05) | 0.374 | 57 (83.9) | 11 (16.2) | 1.27 (0.49, 3.27) | 0.626 |
| HCQ | 16 (64.0) | 9 (36.0) | 1.22 (0.41, 3.63) | 0.722 | 13 (59.1) | 9 (40.9) | 0.59 (0.16, 2.18) | 0.431 | 3 (100.00 | 0 | – | – |
| Favipiravir | 140 (80.0) | 35 (20.0) | 1.51 (0.93, 2.47) | 0.099 | 7 (53.8) | 6 (46.2) | 1.53 (0.74, 3.17) | 0.253 | 100 (84.0) | 19 (16.0) | 1.47 (0.71, 3.03) | 0.297 |
| Methylpred | 110 (90.9) | 11 (9.1) | 1.40 (0.65, 3.02) | 0.395 | 40 (71.4) | 16 (28.6) | 1.15 (0.41,3.28) | 0.789 | 77 (95.1) | 4 (4.9) | 1.38 (0.38, 4.99) | 0.620 |
| Clexane | 220 (80.6) | 53 (19.4) | 1.35 (0.86, 2.11) | 0.190 | 33 (82.5) | 7 (17.5) | 1.02 (0.51, 2.03) | 0.958 | 167 (83.5) | 33 (16.5) | 1.39 (0.69, 2.78) | 0.358 |
| Tocilizumab | 97 (88.2) | 13 (11.8) | 1.25 (0.62, 2.51) | 0.528 | 53 (72.6) | 20 (27.4) | 1.07 (0.46, 2.51) | 0.879 | 59 (95.2) | 3 (4.8) | 1.30 (0.32, 5.37) | 0.717 |
| Dexamethasone | 379 (83.3) | 76 (16.7) | 1.38 (0.78, 2.44) | 0.266 | 38 (79.2) | 10 (20.8) | 0.72 (0.32, 1.61) | 0.420 | 260 (89.3) | 31 (10.7) | 2.07 (0.86, 5.01) | 0.106 |
| Comorbidities | | | | | | | | | | | | |
| CLD | 93 (80.9) | 22 (19.1) | 19.31 (2.38, 156.53) | 0.006 | 119 (72.6) | 45 (27.4) | 12.61 (1.33, 119.46) | 0.027 | 65 (91.5) | 6 (8.5) | – | 0.998 |
| NAFLD | 87 (73.1) | 21 (19.4) | 0.22 (0.03, 1.83) | 0.160 | 58 (77.3) | 17 (22.7) | 0.41 (0.04, 4.10) | 0.445 | 53 (89.8) | 6 (10.2) | – | 0.999 |
| DM | 285 (73.5) | 103 (26.5) | 0.94 (0.66, 1.36) | 0.751 | 51 (76.1) | 16 (23.9) | 0.88 (0.52, 1.44) | 0.599 | 172 (82.3) | 37 (17.7) | 0.93 (0.53, 1.64) | 0.813 |
| HTN | 323 (73.2) | 118 (26.8) | 1.64 (1.16, 2.33) | 0.005 | 113 (63.1) | 66 (36.9) | 1.42 (0.89, 2.26) | 0.143 | 195 (82.6) | 41 (17.4) | 2.13 (1.21, 3.75) | 0.009 |
| Other Comorbidities* | 731(67.6) | 84 (32.4) | 0.91 (0.64, 1.31) | 0.628 | 128 (62.4) | 77 (37.6) | 1.03 (0.64, 1.67) | 0.901 | 94 (76.4) | 29 (23.6) | 0.87 (0.48, 1.56) | 0.631 |
| Ventilator | | | | | | | | | | | | |
| 426 (47.8) | 466 (52.2) | 1.13 (0.69, 2.17) | 0.710 | 65 (58.6) | 46 (41.4) | 0.75 (0.26, 2.15) | 0.589 | 212 (91.8) | 19 (8.2) | 1.33 (0.54, 3.27) | 0.540 |
| ICU admission | | | | | | | | | | | | |
| 154 (89.5) | 18 (10.5) | 1.71 (0.86, 3.38) | 0.124 | 93 (75.6) | 30 (24.4) | 2.69 (1.01, 6.70) | 0.033 | 85 (96.6) | 3 (3.4) | 2.85 (0.70, 11.55) | 0.143 |
| Death | | | | | | | | | | | | |
| 51 (83.6) | 10 (16.4) | 1.04 (0.44, 2.43) | 0.937 | 69 (82.1) | 15 (17.9) | 1.02 (0.34, 3.09) | 0.966 | 30 (90.9) | 3 (9.1) | 0.94 (0.21, 4.16) | 0.934 |
Table 3Multivarite analysis of COVID-19 patients with liver injury (n = 332)
| Characteristics | Overall
| Cohort A
| Cohort B
|
|---|
Liver injury
| AOR (95% CI) | p-value | Liver injury
| AOR (95% CI) | p-value | Liver injury
| AOR (95% CI) | p-value |
|---|
| Yes | No | Yes | No | Yes | No |
|---|
| Age in years | | | | | | | | | | | | |
| ≤ 60 | 193 (22.8) | 653 (77.2) | 1.16 (0.81, 1.66) | 0.421 | 62 (12.2) | 446 (87.8) | 0.66 (0.37, 1.19) | 0.166 | 131 (38.8) | 207 (61.2) | 1.51 (0.92, 2.47) | 0.102 |
| >60 | 139 (34.8) | 261 (65.3) | – | – | 57 (30.2) | 132 (69.8) | – | – | 82 (38.9) | 129 (61.1) | – | – |
| Sex | | | | | | | | | | | | |
| Female | 122 (19.4) | 506 (80.6) | – | – | 39 (10.5) | 333 (89.5) | – | – | 83 (32.4) | 173 (67.6) | – | – |
| Male | 210 (34.0) | 408 (66.0) | 1.79 (1.33, 2.41) | <0.001 | 80 (24.6) | 245 (75.4) | 2.45 (1.52, 3.95) | <0.001 | 130 (44.4) | 163 (55.6) | 1.53 (1.01, 2.32) | 0.048 |
| Ethnicity | | | | | | | | | | | | |
| Malay | 197 (25.3) | 582 (74.7) | – | – | 64 (15.0) | 363 (85.0) | – | – | 133 (37.8) | 219 (62.2) | – | – |
| Chinese | 78 (33.5) | 155 (66.5) | 0.83 (0.56, 1.21) | 0.331 | 29 (23.0) | 97 (77.0) | 0.64 (0.34, 1.21) | 0.166 | 49 (45.8) | 58 (54.2) | 0.93 (0.55, 1.58) | 0.792 |
| Indian | 43 (22.8) | 146 (77.2) | 0.61 (0.39, 0.95) | 0.029 | 19 (17.0) | 93 (83.0) | 0.80 (0.42, 1.52) | 0.494 | 24 (31.2) | 53 (68.8) | 0.54 (0.29, 1.03) | 0.064 |
| Others | 14 (31.1) | 31 (77.2) | 1.50 (0.71, 3.20) | 0.292 | 7 (21.9) | 25 (78.1) | 1.52 (0.55, 4.21) | 0.423 | 7 (53.8) | 6 (46.2) | 1.48 (0.41, 5.33) | 0.547 |
| COVID-19 severity | | | | | | | | | | | | |
| Mild | 62 (9.6) | 587 (90.4) | – | .– | 35 (7.8) | 415 (92.2) | – | – | 27 (13.6) | 172 (86.4) | – | – |
| Moderate | 56 (28.0) | 144 (72.0) | 3.03 (1.87, 4.92) | 0.000 | 26 (24.1) | 82 (75.9) | 2.13 (1.04, 4.33) | 0.038 | 30 (32.6) | 62 (67.4) | 3.26 (1.59, 6.70) | 0.001 |
| Severe | 214 (53.9) | 183 (46.1) | 4.56 (2.27, 9.18) | 0.000 | 58 (41.7) | 81 (58.3) | 3.48 (1.09, 11.09) | 0.035 | 156 (60.5) | 102 (39.5) | 4.71 (1.82, 12.19) | 0.001 |
| Treatment | | | | | | | | | | | | |
| 257 (43.8) | 330 (56.2) | 1.29 (0.72, 2.31) | 0.390 | 80 (34.6) | 151 (65.4) | 3.74 (1.57, 8.93) | 0.003 | 177 (49.7) | 179 (50.3) | 0.47 (0.19, 1.19) | 0.111 |
| Types of treatment | | | | | | | | | | | | |
| NSAID | 36 (44.4) | 45 (55.6) | 0.99 (0.57, 1.72) | 0.970 | 1 (7.7) | 12 (92.3) | 0.12 (0.01, 1.14) | 0.065 | 35 (51.5) | 33 (48.5) | 1.03 (0.52, 2.01) | 0.939 |
| HCQ | 10 (40.0) | 60.0 (60.0) | 1.57 (0.56–4.40) | 0.389 | 7 (31.8) | 15 (68.2) | 0.66 (0.17–2.58) | 0.562 | 3 (100.0) | – | 0.99 (−) | – |
| Favipiravir | 70 (40.0) | 105 (60.0) | 0.91 (0.61, 1.35) | 0.627 | 15 (26.8) | 41 (73.2) | 0.76 (0.36, 1.59) | 0.460 | 55 (46.2) | 64 (53.8) | 0.89 (0.53, 1.48) | 0.649 |
| Methylpred | 74 (61.2) | 47 (38.8) | 1.39 (0.85, 2.28) | 0.196 | 18 (45.0) | 22 (55.0) | 0.95 (0.41, 2.21) | 0.903 | 56 (69.1) | 25 (30.9) | 1.64 (0.82, 3.29) | 0.163 |
| Clexane | 128 (46.9) | 145 (53.1) | 1.17 (0.80, 1.69) | 0.420 | 24 (32.9) | 49 (67.1) | 0.88 (0.41, 2.21) | 0.716 | 104 (52.0) | 96 (48.0) | 1.30 (0.79, 2.15) | 0.300 |
| Tocilizumab | 59 ( 53.6) | 51 (46.4) | 0.93 (0.31, 2.84) | 0.767 | 21 (43.8) | 27 (56.3) | 1.02 (0.47, 2.21) | 0.954 | 38 (61.3) | 24 (38.7) | 0.77 (0.37, 1.58) | 0.477 |
| Dexamethasone | 223 (49.0) | 232 (51.0) | 1.10 (0.63, 1.92) | 0.735 | 60 (36.6) | 104 (63.4) | 0.59 (0.25, 1.37) | 0.216 | 163 (56.0) | 128 (44.0) | 1.60 (0.68, 3.76) | 0.286 |
| Comorbidities | | | | | | | | | | | | |
| Chronic liver disease | 51 (44.3) | 64 (55.7) | 2.93 (1.04, 8.28) | .043 | 20 (26.7) | 55 (73.3) | 7.32 (1.38, 38.84) | .019 | 46 (64.8) | 25 (35.2) | 1.59 (0.43, 5.96) | .489 |
| NAFLD | 46 (42.6) | 62 (57.4) | 0.94 (0.31, 2.84) | .917 | 16 (23.9) | 51 (76.1) | 0.26 (0.05, 1.56) | .141 | 40 (67.8) | 19 (32.3) | 2.44 (0.58, 10.20) | .221 |
| Diabetes mellitus | 143 (36.9) | 245 (63.1) | 0.80 (0.67, 1.14) | .214 | 40 (22.3) | 139 (77.7) | 0.44 (0.25, 0.79) | .006 | 103 (49.3) | 106 (50.7) | 1.12 (0.70, 1.78) | .645 |
| Hypertension | 159 (36.1) | 282 (63.9) | 1.08 (0.77, 1.52) | .649 | 152 (74.1) | 53 (25.9) | 1.31 (0.77, 2.24) | .327 | 106 (44.9) | 130 (44.9) | 0.93 (0.57, 1.50) | .754 |
| Other Comorbidities* | 76 (34.1) | 147 (65.9) | 1.01 (0.70, 1.44) | .974 | 40 (29.4) | 96 (70.6) | 1.38 (0.788, 2.44) | .269 | 50 (40.7) | 73 (59.3) | 0.76 (0.46, 1.26) | .288 |
| Ventilator | | | | | | | | | | | | |
| 193 (54.5) | 161 (45.5) | 1.51 (0.89, 2.56) | .129 | 50 (40.7) | 73 (59.3) | 0.73 (0.26, 2.02) | .540 | 143 (61.9) | 88 (38.1) | 2.10 (1.10, 4.02) | .025 |
| ICU admission | | | | | | | | | | | | |
| 104 (60.5) | 68 (39.5) | 1.60 (0.99, 2.58) | .054 | 41 (48.8) | 43 (51.2) | 2.75 (1.23, 6.19) | .014 | 63 (71.6) | 25 (28.4) | 1.72 (0.84, 3.52) | .136 |
| Death | | | | | | | | | | | | |
| 34 (55.7) | 27 (44.3) | 1.25 (0.66, 2.38) | .489 | 11 (39.3) | 17 (60.7) | 0.83 (0.29, 2.37) | .727 | 23 (69.7) | 10 (30.3) | 1.84 (0.75, 4.55) | .184 |
Differences between cohort A and cohort B
In cohort B, patients were significantly older and more had diabetes mellitus (DM) and HTN compared to cohort A. The majority of patients in cohort A had mild disease (64.6%), whereas cohort B had significantly more patients with severe disease (47%) and required specific treatment, mechanical ventilation, and ICU admission. Notably, both ALB (71.6% vs. 48.5%, p < 0.001) and LI (38.8% vs. 17.1%, p < 0.001) were more prevalent in cohort B patients. The mean value of ALT and AST is consistently higher in cohort B across all severity of disease for both ALB and LI. Cohort B patients with ALB had significantly higher mean values of ALT, AST, and GGT, whereas ALT and GGT were significantly higher in those with LI (Fig. 1). The mean values of ALT and AST were consistently higher in cohort B across all severities of disease for both ALB and LI (Figs. 2 and 3).
Multivariate analysis of cohort A found a significantly higher risk of developing both ALB and LI in male patients, those with underlying CLD, who received specific disease-modulating treatment, or who required ICU admission. Additionally, patients in cohort A with moderate or severe disease had a significantly higher risk of developing LI, with a lesser risk in DM patients (Tables 2 and 3). In cohort B, multivariate analysis identified male sex, moderate or severe disease, and HTN as factors that increased the risk of ALB. Male sex, moderate or severe disease, and requiring mechanical ventilation increased the risk of LI (Tables 2 and 3).
Discussion
From our data, the incidence of mild liver dysfunction in the form of ALB was common (58.7%) among COVID-19 patients. Worldwide, the incidence of mild liver dysfunction ranges from 11.0% to 76.3%.12,14,16,19–27 Lack of a standard definition of liver dysfunction and heterogeneity of the study population probably contribute to the wide variation. Elevated ALT, AST, and GGT had nearly the same incidence, 36.83%, 34.99%, and 34.91% respectively, which mirrors the findings in other studies that included GGT in the analysis.25,27 Our data also showed that severe liver dysfunction, represented by LI (26.6%) was less common, and liver failure (1.1%) was very rare in these hospitalized COVID-19 patients. A similar frequency of severe liver dysfunction was reported by Cai et al.16 and Phipps et al.19 at 21.5% and 27.4%, respectively. Clinical characteristics that were significantly associated with LI were male sex, moderate and severe COVID-19 infection, and underlying CLD.
Several possibilities for liver dysfunction in COVID-19 infection have been suggested. Firstly, SARS-COV-2 virus has been shown to have a direct cytopathic effect on hepatocytes and cholangiocytes.11,17,28,29 This is supported by presence of SARS-CoV 2-interacting host receptors such as angiotensin-converting enzyme 2, transmembrane serine protease 2, and paired basic amino acid cleaving enzyme, which is expressed at varying levels in hepatocytes, cholangiocytes, and periportal liver sinusoidal endothelial cells.30,31 Also, as enterocytes are able to bind angiotensin-converting enzyme 2, there is a high possibility of portal vein viremia leading to an increased likelihood of a direct cytopathic effect.30 Furthermore, analysis of liver samples from two deceased COVID-19 patients with elevated liver enzymes demonstrated the presence of intact viral particles in the cytoplasm of hepatocytes.32 Secondly, pre-existing CLD may be a contributing factor for liver dysfunction in COVID-19 infection. CLD in COVID-19 patients is a well-established risk factor for developing severe liver dysfunction,16,33–35 which is also reflected in our study findings (Tables 2 and 3). However, two other studies reported that CLD was not a significant risk factor for developing liver dysfunction.19,27 This could be attributed to the inclusion of nonalcoholic fatty liver disease which represents the majority of CLD patients and the low prevalence of CLD among the study patients17 as well as the small number of patients in the study.27 Thirdly, drug-induced LI has been postulated to cause abnormal liver function in COVID-19 patients. The use of disease-modulating pharmacotherapy such as hydroxychloroquine (HCQ),17 lopinavir/ritonavir,16 or even tocilizumab11,28,29 in moderate and severe COVID-19 infection has been widely reported to cause drug-induced LI. Cohort A patients, who were treated with disease-modulating drugs, were found to have significantly increased occurrences of ALB and LI, but this trend was not observed in cohort B patients. One possible explanation for this would be the use of HCQ in our center during the early wave of COVID-19 infection which fell out of favor thereafter. Hundt et al.20 reported significant liver dysfunction with the use of HCQ, but another center reported no increase in liver dysfunction associated with HCQ.27
Additional factors that have been postulated to cause liver dysfunction in COVID-19 infection are cytokine storm, ischemic hepatitis, and liver congestion associated with mechanical ventilation.17,28,29 In our study, LI was more frequent in Delta-variant patients who required ventilatory support than in those who did not require it. Mechanical ventilation19,20 and the presence of cytokine storm, in which inflammatory markers such as interleukin-6 and ferritin are elevated have been associated with significant liver dysfunction.19 The presence of these factors usually implies severe COVID-19 infection. This is in line with our data that showed patients with moderate and severe COVID-19 infection were more likely to develop ALB and LI. Multiple studies have demonstrated that the severity of liver dysfunction was proportionate to the severity of COVID-19 infection.19,20,26,27
Apart from the severity of illness and presence of CLD, male sex increased the likelihood of developing ALB and LI. Other studies have reported significantly higher liver dysfunction in men.20,27 Previous studies have shown that men tended to have a more severe course of illness than women. This was possibly due to several factors such as the difference in sex-related immunological response driven by sex hormones and the X chromosome as well as a high expression of coronavirus receptors (angiotensin-converting enzyme 2) in men.36,37 Less healthy lifestyle options such as higher levels of smoking and drinking among men than in women may also be a contributing factor.
Only a handful of studies have analyzed the clinical characteristics of the Delta variant. Budhiraja et al.,38 who reviewed nearly 20,000 COVID-19 patients, found that almost 40% of Delta-variant patients had a severe course of disease with significantly increased need of ICU admission, oxygen support, and use of remdesivir, steroids, intravenous immunoglobulin, and enoxaparin. Interestingly, a significant difference in age was not observed, although presence of the Delta variant was notably higher in male patients and in those with DM and HTN.38 Al Bahrani et al.,39 who analyzed 619 patients, reported those infected with the Delta variant were significantly older but reported a similar need of ICU admission, use of mechanical ventilation, and use of methylprednisolone for both groups of patients. The mortality rate among Delta-variant patients was lower than the rates in patients with other VoCs.
The frequency of ALB and LI was higher among patients with Delta variant COVID-19 infection (cohort B) compared with predecessor strains (cohort A). In particular, transaminitis was higher in patients infected with Delta variant compared with predecessor strains across all categories of severity for both ALB and LI. Al Bahrani et al.39 reported similar findings, where ALT and AST were significantly higher in patients infected with the Delta variant.39 Differences in admission criteria that led to differences in the baseline demographics of the cohorts may have contributed to significantly higher ALB and LI rates in patients with mild disease in cohort B than in cohort A, as this would have excluded a proportion of asymptomatic Delta variant patients who most likely would have had no or very slight liver derangement. However, the reason for the higher proportion of ALB and LI among Delta variant patients with moderate and severe disease is not well understood. A possible explanation is that Delta variant has a higher affinity to hepatocytes and cholangiocytes which results in a higher incidence of hepatitis and cholestasis. The Delta variant has a higher cell affinity with better cell fusion and enhanced cell entry.4,40 Arora et al.40 demonstrated the Delta variant to have enhanced lung and colon cell entry, resulting in more tissue damage and higher virulence than previous variants. This could likely result in a higher burden of portal vein viraemia, further increasing the risk of damage to hepatocytes and cholangiocytes.
This study has some limitations. As the study population is based on convenience sampling, some of the patients with mild disease and relatively short stays with normal LFTs on admission do not have a repeat blood test. Hence, there could be a small proportion of patients who developed liver dysfunction during hospitalization that was not captured in the data. Furthermore, for this same reason, baseline and peak abnormalities of LFT could not be compared. Differences in admission criteria for both cohorts also led to sampling bias among patients with mild disease in this study. As this is a single-center study in an urban area, these data may not be representative of patients in a semi-urban or rural setting. Additionally, vaccination history, which may contribute to liver dysfunction, was not included due to a lack of data.