Assessment of abnormal liver function tests and associated factors among COVID-19-infected patients in Addis Ababa, Ethiopia, 2022: a facility-based comparative cross-sectional study

Abstract

Objective

Liver function test (LFT) abnormalities are higher in patients with severe COVID-19. Most of the studies on this theme were conducted in foreign nations, and the association with LFT abnormalities was not sufficiently addressed in the study areas. Therefore, the current study aimed to investigate the effects of COVID-19 infection on liver function of patients.

Setting

A facility-based comparative cross-sectional study was carried out from 10 April to 15 June 2022, among COVID-19 infected individuals admitted in Eka Kotebe General Hospital and Saint Petrous Specialized Hospitals, Addis Ababa, 2022.

Participants

A total of 284 confirmed COVID-19-positive and COVID-19-negative controls matched by gender and age were included in the present study.

Results

Among SARS-COV-2 positive groups, 63 (44.4%) had one or more LFT abnormalities. The most common elevated level of the LFTs among patients with COVID-19 were gamma-glutamyl transferase (GGT) 50 (35.2%), while the most common lowered level was albumin 58 (40.8%). The mean values of aspartate aminotransferase (AST) (35.4±26.9 vs 22.9±12.6, p<0.001) were significantly different between patients with COVID-19 and the COVID-19-free groups. Being COVID-19-positive was significantly associated with an elevated level of AST (AOR=3.0, 95% CI 1.2 to 7.4) and GGT (AOR=4.55, 95% CI 2.02 to 10.3). Being male was significantly associated with an elevated level of total bilirubin (BILT, AOR=2.41, 95% CI 1.2 to 4.9) and direct bilirubin (BILD, AOR=3.7, 95% CI 1.72 to 8.2), and also severe stage of COVID-19 was associated with hypoalbuminaemia (AOR=3.3, 95% CI 1.4 to 7.9). SARS-COV-2 infection was independently associated with LFT abnormality.

Conclusion

Patients with COVID-19 had decreased albumin levels, and elevated AST, GGT, BILT and BILD levels.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The comparison methodology we adopted is one of the strengths of the study.

• This study was conducted at two different facilities. In this study consistent definitions were used to reduce bias among readers.

• The study has a cross-sectional study design and thus, it is not possible to show a case-effect relationship and the use of convenient sampling technique (lack of generalisability due to vulnerability of selection bias).

• On the present study screening tests were not performed in recruting participants and to handle confounding comorbidities.

• Due to unavailability of these assays and cost, serum alkaline phosphatase (ALP) was analysed without identifying bone-specific ALP.

Introduction

Coronaviruses are enveloped viruses with spikes around their outer structure, containing a constructive single-stranded RNA with a full length of approximately 32 kb acting as mRNA during the replicated polyprotein translation.¹ SARS-CoV-2 causes the infectious disease known as COVID-19.² These viruses are members of the subfamily Coronaviridae and the order Nidovirales.³ On 11 March 2020, the WHO declared the COVID-19 outbreak a pandemic.⁴ COVID-19 symptoms can range from mild to severe and are mainly characterised by fever, dry cough, dyspnoea, headache, sore throat and rhinorrhoea.⁵ Past infections, clinical demonstrations and clinical testing are all used to screen for coronavirus; a swab test for viral nucleic acid, a CT scan, an antigen rapid diagnostic test (RDT), and a blood culture can be used for diagnosis. Real-time PCR is the ‘gold standard’ method for detecting human coronavirus.¹

The liver plays several important roles and acts as a hub in our body. The most important functions of the liver are its metabolic, synthetic and excretory functions. Liver function tests (LFTs) are a set of biochemical tests analysed to evaluate the specific function of liver cells. The common liver function tests include total bilirubin (BILT), direct bilirubin (BILD), serum alanine aminotransferase (ALT), and aspartate aminotransferase (AST) which under normal physiological state present inside the liver are released into the circulation after hepatocyte injury. Serum alkaline phosphatase (ALP) and gamma-glutamyl transferase (GGT) are located in the cholangiocytes and represent their injury, and serum albumin is a marker of the synthetic functions of the liver.⁶

In addition to respiratory failure, many SARS-CoV-2 deaths have been caused by multiple organ dysfunction syndromes (MODS). This could be due to the presence of ACE-2, which is the principal SARS-CoV-2 receptor, across various organs.⁷ The liver is one of the most frequently impaired organs, besides the lungs. This is a potential target for SARS-CoV-2 infection due to its expression of ACE2. ACE2 expression is 20-fold higher in cholangiocytes than in hepatocytes. Cholangiocytes and hepatic endothelial cells have been proposed to be the target cells for SARS-CoV-2.⁸ The entry of SARS-CoV-2 into the liver causes COVID-19-associated liver injury and, also, the elevation of serum aminotransferases in clinical samples.8,10

Multiple mechanisms have been postulated for how the virus causes SARS-CoV-2-related liver injury, including the virus’s direct cytopathic effect on cholangiocytes, immune-mediated damage, hypoxic sepsis and medication interactions.¹¹ Cholangiocytes play a role in many aspects of liver physiology, including regeneration and adaptive immune response mechanisms, and their function can be disrupted, resulting in hepatobiliary injury.¹² Cholangiocytes express not only the ACE-2 receptor but also the transmembrane serine protease 2, which cleaves the S protein of the virus before it enters cells, providing the basis for cholangiocytes being highly vulnerable to SARS-CoV-2 damage.¹³ The major pathophysiological features of liver injury include enlargement of liver volume, liver cell focal necrosis with neutrophil infiltration, hepatocyte steatosis, lobular and portal inflammation, hepatic sinus congestion, microthrombosis, and high filling of the gall bladder.¹⁴ Finally, this results in abnormal LFTs, which are typically found in patients with COVID-19. They show significantly higher levels of hepatocellular and cholangiocellular damage, indicating liver function biochemical indicators like serum ALT, AST, GGT and ALP.¹⁴

The COVID-19 epidemiology has changed since its first report in China on 31 December 2019. As the WHO report shows globally, as of 28 August 2022, over 598 million confirmed cases and over 6.4 million deaths had been reported.¹⁵ COVID-19 was a serious threat to public health worldwide, with a global mortality rate of 5.15% as of 24 June 2020.16,18 Similarly, Egypt reported the first COVID-19 case in Africa on 14 February 2020, and Algeria on 25 February 2020.¹⁹ In Ethiopia, from 13 March 2020 to 20 September 2022, there have been 493 456 confirmed cases of COVID-19, with 7572 deaths.^{20 21} Critically ill individuals might suffer from septic shock, severe metabolic acidosis, coagulopathy, thrombocytopenia, arrhythmias and multiple organ failure syndromes.²²

In 54% of patients with COVID-19, ACE2 expression was found to be high in bile duct cells, as shown by increased GGT levels. COVID-19 is frequently associated with mild liver function abnormalities, particularly in individuals with severe or critical COVID-19 who are older. The abnormalities in LFTs are very high in patients with severe COVID-19, with a rate of 14%–53% previously recorded, most typically in a hepatocellular pattern.^{18 23 24}

The mortality rate in patients with cirrhosis and COVID-19 was similar to that in patients with cirrhosis alone (30% vs 20%), and higher than that in patients with COVID-19 alone (30% vs 13%). In addition, about 2%–11% of patients with COVID-19 had underlying chronic liver disease and patients with chronic hepatitis B co-infected with the SARS virus are more likely to develop severe inflammation of the liver.^{25 26} A higher proportion of enzyme elevation was noted in patients receiving lopinavir/ritonavir therapy (56.1% vs 25%), and the use of lopinavir/ritonavir was also found to lead to increased odds of liver injury.^{26 27} Comorbidities like metabolic-associated fatty liver disease, obesity and diabetes were present in nearly 80% of the patients. The presence of obesity is known to increase the risk of liver injury. A higher incidence of acute liver injury has also been reported in diabetics without liver disease.²⁸

The majority of studies on COVID-19, which is currently a pandemic disease and a major problem in the world, were carried out in other countries. As most studies show, the severity or prevalence of abnormal LFTs among patients with COVID-19 was affected by geography and ethnicity, and the link between liver failures was not effectively addressed in our country, Ethiopia. As a result, the goal of this study is to show the relationship between abnormal LFTs and COVID-19 disease in the study area. Finally, the current study will show the impacts of COVID-19 on liver functioning by calculating the mean level of the LFTs as compared with healthy controls. In addition, this information will be crucial in enhancing patient care and preventing or minimising hepatic dysfunctions in patients infected with SARS-CoV-2.

Methods

Study area, design and period

A facility-based comparative cross-sectional study was carried out in COVID-19 diagnostic and treatment centre hospitals in the city of Addis Ababa from 10 April to 15 June 2022. ‘Eka Kotebe general hospital’ and ‘Saint Petrous specialized hospitals’ were included in the current study.

Study populations

The current study comprises a total of 284 participants enrolled in two comparable study groups. The first group comprises patients with confirmed COVID-19 and the other group comprises gender-matched and age-matched confirmed COVID-19-negative healthy controls.

Eligibility criteria

Inclusion criteria

All COVID-19-confirmed individuals admitted to Eka Kotebe General Hospital and Saint Petrous Specialized Hospital and who were willing to voluntarily participate in the study were included. Past medical history and current illnesses, including pre-existing liver disease, a history of diabetes mellitus, pre-existing hypertension, a history of hepatitis, a history of HIV/AIDS, a history of heart disease and a history of medication were obtained by reviewing medical records and face-to-face interviewing (online supplemental file 1). All screened COVID-19-free individuals who visited Eka Kotebe General Hospital and Saint Petrous Specialized Hospital and who were willing to voluntarily participate in the study and those who had no previous history of SARS-CoV-2 infection were included as a healthy control group (online supplemental file 2).

Exclusion criteria

For the COVID-19 study group

Individuals who declined to give consent for participating, those who had liver transplantation, patients with COVID-19 with cancer who started taking medication, patients who took medication such as anti-TB drugs (pyrazinamide), antivirals (niverapine), antiepileptic drugs such as carbamazepine and clozapine, aspirin and antidepressants, antipsychotic drugs²⁹ and acetaminophen,³⁰ and pregnant women³¹ were excluded from the study (online supplemental file 1).

For healthy controls

Individuals who declined to give consent for participating, those who had liver transplantation, those who were diagnosed with cancer, and pregnant women were excluded. These tasks were accomplished by preparing a checklist for exclusion (online supplemental file 2).

Sample size and sampling technique

The sample size was determined by using a double population proportion formula. Study subjects were recruited using a convenient sampling technique. The sample size was calculated using Epi Info using information from a study conducted at NewYork-Presbyterian Hospital.³² Multiple sample size calculations were performed and we took the maximum sample size for each outcome variables. The final sample size was 284 suspected patients with COVID-19 in both groups enrolled (142 confirmed and 142 healthy control groups).

Study variables

Dependent variable

LFT panel.

Independent predictor variables

• Sociodemographic characteristics (age, sex, educational status, occupation, marital status, income).

• Behavioural characteristics: smoking habit, alcohol consumption, chat chewing, physical activity.

• Anthropometric and clinical characteristics: body mass index (BMI), blood pressure (BP), pre-existing diabetes mellitus, history of hypertension, nutritional status, history of hepatitis, pre-existing kidney disease, pre-existing liver disease, pre-existing cardiovascular disease, history of vaccinations, individual who uses drugs and stage of COVID-19 diseases. These patients’ clinical characteristics were taken from their medical chart and they were asked about their previous exposures using a semistructured questionnaire (onlinesupplemental files 1 2).

Data collection procedure

First, written informed consent and assent were obtained from the study participants. Then sociodemographic characteristics and participants’ clinical data were collected by trained nurses using a pretested, semistructured questionnaire through a face-to-face interview and medical records reviewed during their hospital visit (onlinesupplemental files 1 2).

Anthropometric measurements such as BP and BMI, as well as physical activity, were collected by trained nurses. Height was measured to the nearest 0.5 cm using a stadiometer and weight was recorded to the nearest 0.1 kg with the patient being barefoot and wearing light clothes using a balance. BMI was calculated as weight divided by height squared (kg/m²). BP was measured in a sitting position on the right arm using an analogue sphygmomanometer and stethoscope. Two readings were taken 5 min apart, and the average was taken as the final BP reading and the status was defined according to the American Society of Hypertension’s definition of hypertension.³³

All COVID-19 protocols were followed in all the data collection and lab analysis procedures, and all technical operations were carried out in such a way that the creation of aerosols and droplets was minimised. Personal protective equipment was used. All laboratory employees and nurse professionals who were collecting blood specimens and sociodemographic data collectors used appropriate disposable gloves and hand sanitiser as established by a rigorous risk assessment. After that, a suitable place for blood sample collection, a medial vein, was chosen, and 70% of alcohol was used. A needle and syringe were used to collect 5 mL of venous whole blood sample and then transfer it into the appropriate serum separator tube.³⁴ After collection, the blood was allowed to clot by leaving it undisturbed at room temperature for about 15–30 min, centrifuged at 3000 revolutions per minute to separate serum from the blood components,³⁵ and transferred into a neck tube. RDT and PCR tests were performed for the recruitment of the healthy controls. To be analysed in a chemistry lab, the serum was deep frozen to −20°C, labelled as a high-risk biological sample, and then transported to Gondar Compressive Specialized Hospital using a triple package (an icebox). All these procedures were done by the laboratory technologist and principal investigator by applying standard operating procedures. Finally, the serum sample was analysed using the DxC 700 AU auto analyzer (Beckman Coulter, USA).

Laboratory procedures

LFT was performed on the serum using the Beckman Coulter DxC 700 Automated Chemistry Analyzer; the DxC 700 AU clinical chemistry analyzer is the latest innovation manufactured by Beckman Coulter, a Danaher Corporation company in Brea, California, USA. Enzymatic LFT results were reported as IU/L, and the other LFT results were reported in mg/dL.

Serum AST activity was measured by the method in which AST catalyses the transamination of aspartate and α-oxoglutarate, forming L-glutamate and oxalacetate. The oxalacetate is then reduced to L-malate by malate dehydrogenase, while NADH is simultaneously converted to NAD+. The decrease in absorbance due to the consumption of NADH is measured at 340 nm and is proportional to the AST activity in the sample.³⁶

Serum ALT activity was measured by the following method: ALT transfers the amino group from alanine to α-oxoglutarate to form pyruvate and glutamate. The pyruvate enters a lactate dehydrogenase-catalysed reaction with NADH to produce lactate and NAD+. The decrease in absorbance due to the consumption of NADH is measured at 340 nm and is proportional to the ALT activity in the sample.³⁶

Serum ALP activity is determined by measuring the rate of conversion of p-nitrophenyl phosphate (pNPP) in the presence of 2-amino-2-methyl-1-propanol at pH 10.4. The rate of change in absorbance due to the formation of pNPP is measured at 410/480 nm and is directly proportional to the ALP activity in the sample.

BILT was measured using the classical method. A stabilised diazonium salt, 3, 5-dichlorophenyldiazonium tetrafluoroborate, reacts with bilirubin to form azobilirubin, which absorbs at 570/660 nm. Caffeine and a surfactant are used as reaction accelerators. The absorbance at 570/660 nm is proportional to the bilirubin concentration in the sample. A separate serum blank is performed to eliminate absorbance from the serum and interference.³⁶

BILD was measured by using a variation of the classical method. Direct (conjugated) bilirubin is coupled with a diazonium salt of 3, 5-dichloroaniline in an acid medium to form azobilirubin. The BILD in serum is directly proportional to the colour development of azobilirubin which is measured bichromatically at 570/660 nm.

GGT was measured enzymatically because GGT catalyses the transfer of the gamma-glutamyl group from the substrate, gamma-glutamyl-3-carboxy-4-nitroanilide, to glycylglycine, yielding 5-amino-2-nitrobenzoate. The change in absorbance at 410/480 nm is due to the formation of 5-amino-2-nitrobenzoate and is directly proportional to the GGT activity in the sample.

Total protein was measured by the method in which cupric ions in an alkaline solution react with proteins and polypeptides containing at least two peptide bonds to produce a violet-coloured complex. The absorbance of the complex at 540/660 nm is directly proportional to the concentration of protein in the sample. Serum albumin was measured at pH 4.2. Bromocresol green reacts with albumin to form an intense green complex. The absorbance of the albumin bromocresol green (BCG) complex is measured bichromatically (600/800 nm) and is proportional to the albumin concentration in the sample.

Operational definition

LFT abnormality

It was defined as any parameter of the liver enzyme panel greater than the upper limit of the normal reference range. The values of ALT>40 U/L, AST>43 U/L, GGT>65 U/L, ALP>93 U/L, albumin<3.5 mg/dL, BILT>1.1 mg/dL, BILD>0.2 mg/dL were considered as abnormal.³⁷

Data quality control

Prior to actual data collection, the prepared questionnaires were translated from English to Amharic and pretested at ‘Saint Paul’ hospital on about 5% of the sample size (15 individuals) to assure their accuracy and consistency. Training was given to data collectors about the objective and relevance of the study, confidentiality issues, study participants’ rights, consenting, interview techniques, and laboratory test procedures and quality control. The blood sample collection and laboratory tests were performed by well-trained nurses and laboratory technologists. The collected samples were analysed by the Beckman Coulter DxC 700 chemistry analyzer, and their quality was checked by previously documented Amhara regional state, regional laboratory, and EQA sample result feedback reports.

Standard quality control (normal and pathological) protocols were performed and assured before running the participants’ sample to assure the analytical performance, accuracy and functionality of the instrument. Furthermore, the investigator closely followed and frequently checked the sample analysis process to ensure its completeness and consistency.

Data analysis and interpretation

The collected data were entered using EpiData V.4.6.0.0 and transferred to STATA V.14. The normality of data distribution was checked using an histogram. Additionally, multicollinearity of the variables was checked by variance inflation factors. The cleaned data were statistically analysed using the χ² test, independent t-test, Mann-Whitney test and logistic regression to measure the statistical association between LFT abnormalities and their associated risk factors among patients with COVID-19. Hosmer-Lemeshow tests were performed. Bivariate analysis was done to ascertain any significant association of the associated variables with abnormal LFTs among individuals with COVID-19. Any variables with a value of p<0.25 in bivariate analysis were entered in multivariable logistic regression analysis. Descriptive statistics were performed, and after that, continuous variables were reported as mean $\pm$SD and IQR and compared using an independent t-test and Mann-Whitney test. A value of p<0.05 was considered statistically significant, and AOR and 95% CI were used to give an appropriate interpretation of the data and discussion.³⁸ Finally, categorical variables were compared using the χ² test and reported as frequencies and percentages in tables and graphs.

Ethical consideration

SpecializedSpecialisedHospitalInformed consent was obtained from each study participant, and for participants less than 18 years old, assent was obtained from their parents. The confidentiality of the collected information was maintained by not recording any personal identifiers. In addition, the clinical specimen collected during the study period was used for the stated objectives only, and the study participants participated once during the study period.

Patient and public involvement

Confirmed COVID-19-positive individuals, gender-matched and age-matched confirmed COVID-19 negative controls at ‘Eka Kotebe General Hospital’ and ‘Saint Petrous Specialized Hospital’ participated in the study and provided useful information and blood samples for the study. However, they have never been involved in the study design, protocol, data collection tools, reporting, or dissemination of the findings. The findings will be shared with the Addis Ababa city health office as well as published in an open-access international journal.

Results

Sociodemographic characteristics of the study participants

This study included 142 patients infected with SARS-CoV-2 and 142 SARS-CoV-2-free healthy controls with a response rate of 100%. Among the study groups, 185 (65.1%) were male, ranging in age from 6 years to 90 years (mean±SD 45.81±17.84 years). The mean±SD age of the SARS-CoV-2-positive and control groups was 47±17.7 years and 44.5±17.9 years, respectively. The majority of the study participants were urban residents (269 (94.7%)) and 253 (89.0%) were married. 107 (37.6%) study participants had completed primary school, 92 (32.4%) were merchants and 161 (56.7%) had a monthly income >Ethiopian birr 4500 (table 1).

Table 1. Sociodemographic characteristics of the study participants (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022.

Variable	Category	COVID-19 positive, n=142, No (%)	Controls, n=142,No (%)	Total n=284,No (%)	P value
Sex	Male	96 (67.6)	89 (62.7)	185 (65.1)	0.383
	Female	46 (32.4)	53 (37.3)	99 (34.9)
Age	≤20 years	9 (6.3)	11 (7.7)	20 (7.04)	0.896
	21–64 years	105 (73.9)	103 (72.5)	208 (73.2)
	>64 years	28 (19.7)	28 (19.7)	56 (19.72)
Marital status	Single	12 (8.4)	19 (13.4)	31 (21.8)	0.183
	Married	130 (91.8)	123 (86.6)	253 (89.0)
Residence	Rural	11 (7.7)	4 (2.8)	15 (5.2)	0.63
	Urban	131 (92.2)	138 (97.2)	269 (94.7)
Educational status	Unable to read and write	9 (6.3)	14 (9.8)	23 (8.0)	0.025*
	Primary Co	44 (30.9)	63 (44.3)	107 (37.6)
	Secondary Co	28 (19.7)	26 (18.3)	54 (19.0)
	College/university	61 (42.9)	39 (27.5)	100 (35.2)
Occupation	Government	52 (36.6)	38 (27.7)	90 (31.6)	0.095
	Housewife	15 (10.5)	18 (12.6)	33 (11.6)
	Merchant	47 (33.0)	45 (31.7)	92 (32.4)
	Student	4 (2.8)	15 (10.5)	19 (6.6)
	Farmer	8 (5.6)	11 (7.7)	19 (6.6)
	Unemployed	16 (11.3)	15 (10.5)	31 (10.9)

^*Statistically significant.

Clinical and anthropometric characteristics of the study participants

The most prevalent clinical characteristics of the COVID-19-infected group were diastolic hypotension (14.8%); 14.08%, 2.11% and 2.8% had a history of diabetes mellitus, a history of heart disease and pre-existing liver disease, respectively. Of the study participants, 168 (59.1%) had a normal BMI, 80 (28.1%) were overweight and 15 (5.2%) were obese (table 2).

Table 2. Clinical and anthropometric characteristics of study participants (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022.

Variable	Category	COVID-19 positive, n=142, No (%)	Controls,n=142, No (%)	Totaln=284, No (%)	P value
Hypertensive	Yes	16 (11.27)	49 (34.5)	65 (22.9)	<0.001*
	No	126 (88.7)	93 (65.49)	219 (77.11)
History of heart disease	Yes	3 (2.11)	31 (21.8)	34 (11.97)	<0.001*
	No	139 (97.9)	111 (78.17)	250 (88.02)
History of diabetes	Yes	20 (14.08)	17 (11.97)	37 (13.03)	0.597
	No	122 (85.92)	125 (88.03)	247 (86.97)
Pre-existing kidney disease	Yes	14 (9.85)	29 (20.42)	43 (15.14)	0.013*
	No	128 (90.14)	113 (79.6)	241 (84.85)
History of HIV/AIDS	Yes	15 (10.56)	3 (2.11)	18 (6.34)	0.003*
	No	127 (89.44)	139 (97.9)	266 (93.7)
History of hepatitis	Yes	4 (2.82)	6 (4.23)	10 (3.52)	0.520
	No	138 (97.18)	136 (95.8)	274 (96.47)
Pre-existing liver disease	Yes	4 (2.82)	13 9.15)	17 (5.98)	0.024*
	No	138 (97.18)	129 (90.85)	267 (94.01)
Systolic blood pressure	Hypotension	6 (4.22)	4 (2.82)	10 (3.52)	0.001*
	Normotension	117 (82.4)	93 (65.5)	210 (73.9)
	Hypertension	19 (13.4)	45 (31.7)	64 (22.53)
Diastolic blood pressure	Hypotension	21 (14.8)	21 (14.8)	42 (14.8)	0.072*
	Normotension	105 (73.9)	91 (64.1)	196 (69.01)
	Hypertension	16 (11.26)	30 (21.13)	46 (16.2)
Antiviral drugs	Hydroxychloroquine	69 (24.30)	–	69 (24.30)	NA
	Chloroquine	73 (25.7)	–	73 (25.7)
Body mass index	Underweight (<18)	4 (2.8)	17 (11.9)	21 (7.3)	0.01*
	Normal weight (18–24.9)	94 (66.1)	74 (52.1)	168 (59.1)
	Overweight (25-30)	36 (25.3)	44 (30.9)	80 (28.1)
	Obese (30+)	8 (5.6)	7 (4.9)	15 (5.2)

^*Statistically significant.

Behavioural characteristics of the study participants

Of the study participants, 281 (98.94%) were non-smokers, 223 (78.5%) were non-drinkers and 144 (50.7%) did intense physical exercise (table 3).

Table 3. Behavioural characteristics of participants (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022.

Variable	Category	COVID-19 positive,n=142, No (%)	Controls, n=142,No (%)	Total n=284,No (%)	P value
Physical exercise	Inactive	70 (49.3)	54 (38.03)	124 (43.7)	0.056
	Active	72 (50.7)	88 (61.97)	160 (56.33)
Alcohol consumption	Non-drinker	102 (71.8)	121 (85.2)	223 (78.5)	0.006
	Frequent drinker	40 (28.2)	21 (14.8)	61 (21.5)
Smoking habit	Yes	1 (0.7)	2 (1.41)	3 (1.06)	0.562
	No	141 (99.3)	140 (98.6)	281 (98.94)

The prevalence of LFT abnormalities among the study participants

Of the 284 participants in the study, 134/284 (47.2%) had abnormal LFTs. Additionally, among the SARS-CoV-2-positive groups, 82/142 (57.7%) exhibited abnormal LFTs, whereas abnormal LFTs in the control groups were 52/142 (36.6%) (figure 1).

Figure 1. Comparison of liver function test (LFT) abnormalities between the COVID-19 infected group and control group (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022.

Comparison of elevated LFTs between the COVID-19 infected group and control group

The most frequent liver function test abrormality among SARS-CoV-2-positive was elevated GGT, and decreased albumin followed by elevated BILT and BILD (figure 2).

Figure 2. Comparison of elevated liver function tests (LFTs) between the COVID-19 infected group and control group (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022. ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BILD, direct bilirubin; BILT, total bilirubin; GGT, gamma-glutamyl transferase.

The mean comparison of LFTs between COVID-19-infected and COVID-19-free healthy control patients

In this study, the subjects who tested positive for SARS-CoV-2 had higher mean±SD of ALT (24.2±20.4 vs 18.7±12.1, p=0.0031) and BLIT (0.903±0.49 vs 0.64±0.36, p<0.001) than the control group, respectively. On the other hand, the mean±SD level of albumin for the positive group was lower (3.6±0.69, p=0.001) than that of the control group (4.01±0.063, p=0.001). The median and IQR levels of GGT are higher in the SARS-CoV-2-positive group (73 (24±97), p<0.001) than in the healthy controls (22 (14±36), p<0.001) (table 4).

Table 4. The mean comparison of liver function test abnormalities between COVID-19-infected and COVID-19-free patients (n=284 at Addis Ababa COVID-19 diagnostic and treatment centres, Addis Ababa) from March to June 2022.

Liver function tests	Patients with COVID-19 versus healthy controls				Total n=284		P value
	Patients with COVID-19 n=142		Healthy controls n=142		Mean±SD
	Mean±SD		Mean±SD
ALT	24.2±20.4		18.7±12.1		21.5±16.95		0.0031*
AST	35.4±26.9		22.9±12.6		29.2±26.6		<0.001*
GGT	Median 73	IQR (24–97)	Median (22)	IQR (14–36)	Median (40.5)	IQR (17–57.5)	<0.001*
Total bilirubin	0.903±0.49		0.64±0.36		0.77±0.45		<0.001*
Direct bilirubin	0.15±0.093		0.11±0.098		0.13±0.098		<0.001*
ALP	71.5±51.6		64.5±54.97		67.98±53.33		0.1341
Albumin	3.6±0.69		4.01±0.63		3.8±0.696		<0.001*

^*Statistically significant mean difference between two groups.

ALP, alkaline phosphataseALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase;

Multivariate logistic regression analysis of associated factors of LFT abnormalities among the study participants

From multivariate logistic regression analysis, being SARS-CoV-2 positive was significantly associated with elevated level of AST (AOR=3.0, 95% CI 1.2 to 7.4), GGT (AOR=4.55, 95% CI 2.02 to 10.3), BILT (AOR=7.9, 95% CI 3.4 to 18.74) and BILD (AOR=3.72, 95% CI 1.4 to 10.0). Almost all LFTs were not associated with excessive drinking except elevated levels of ALP (AOR=0.25, 95% CI 0.08 to 0.78). Among stages of COVID-19 disease, the severe stage was associated with hypoalbuminaemia (AOR=3.3, 95% CI 1.4 to 7.9). Being male was significantly associated with an elevated level of BILT (AOR=2.41, 95% CI 1.2 to 4.9) and BILD (AOR=3.7, 95% CI 1.72 to 8.2). Previous and current history of diabetes mellitus was associated with an elevated level of BILT (AOR=0.26, 95% CI 0.09 to 0.77) and BILD (AOR=0.28, 95% CI 0.09 to 0.90). In addition, a pre-existing history of heart disease was significantly associated with an elevated level of BILT (AOR=6.8, 95% CI 2.3 to 20.2). And also, diastolic hypotension and hypertension were significantly associated with an elevated level of BILT (AOR=2.9, 95% CI 1.24 to 6.8) (onlinesupplemental tables 1 2).

Discussion

COVID-19 is an infectious disease caused by SARS-CoV-2. In addition to respiratory failure, many SARS-CoV-2 deaths have been caused by MODS. The liver is one of the most frequently impaired organs, besides the lungs. The entry of SARS-CoV-2 into the liver causes COVID-19-associated liver injury and, in addition, the elevation of serum aminotransferases in clinical samples. Multiple mechanisms have been postulated for how the virus causes SARS-CoV-2-related liver injury, including the virus’s direct cytopathic effect on cholangiocytes, immune-mediated damage, hypoxic sepsis and medication interactions. The goal of this study is to show the relationship between abnormal LFTs and COVID-19 disease in the study area. The current study comprises a total of 284 participants enrolled in two comparable study groups.

In this study, the study subjects who tested positive for SARS-CoV-2 had higher mean±SD of ALT and BLIT than the control group. On the other hand, the mean±SD level of albumin for the positive group was lower than that of the control group. Being COVID-19 positive was significantly associated with elevated levels of AST (AOR=3.0, 95% CI 1.2 to 7.4), GGT (AOR=4.55, 95% CI 2.02 to 10.3), BILT (AOR=7.9, 95% CI 3.4 to 18.74) and BILD (AOR=3.72, 95% CI 1.4 to 10.0).

We analysed the serum levels of ALT, AST, GGT, ALP, albumin and BILD, and total serum levels in patients with COVID-19 and a healthy control group. The present study shows that among SARS-CoV-2-positive groups, 57.5% (95% CI 49% to 66%) had abnormal LFTs. This result shows a higher prevalence of LFT abnormalities as compared with a retrospective observational study conducted in Singapore that found an abnormal LFT of 18.4% among 554 study subjects,³⁹ and a study conducted in Egypt that found an abnormal LFT of 44% among 52 study subjects.⁴⁰ This difference might be because undiagnosed comorbidity might increase the prevalence of LFT abnormalities in addition to SARS-CoV-2 infection, and some of the study participants had pre-existing liver disease.⁴¹

Our finding shows that, compared with SARS-CoV-2-negative individuals, SARS-CoV-2-positive individuals were more likely to have a high GGT level of 35.5% (95% CI 27.7% to 43.5%) and hypoalbuminaemia of 40.8% (95% CI 33% to 490.2%). Additionally, they were more likely to have a higher BILT level of 30.9% (95% CI 23% to 390.1%) and BILD level of 33.1% (95% CI 25.8% to 41.3%). Furthermore, an elevated AST level of 22.5% (95% CI 16% to 30.2%), ALT level of 14.7% (95% CI 9.8% to 21.7%) and ALP level of 19.7% (95% CI 13.9% to 27%) were also shown in our study. The retrospective study from Singapore showed a higher level of AST, ALT and ALP in 78 patients (76.5%), 75 patients (75.5%) and 8 patients (7.8%), respectively, than our results. These results showed a higher prevalence of AST and ALT levels than our result but a lower level of ALP and bilirubin than our findings.³⁹ In addition, these results contrast with a retrospective study done in a US hospitals’ network (1827 study participants), which reveals lower ALP and BILT levels than our result but higher AST and ALT, and reduced albumin than our results.⁴² This difference might be due to the study design and study participants’ age differences (specifically, in our study, some young people were included) and COVID-19 patient disease stage differences, because in our study most of the patients were in mild and moderate stages, which may decrease the prevalence. Serum ALT and AST were elevated more frequently and to a greater extent in patients with severe COVID-19 compared with those with mild disease.⁴³ Whereas our result is in line with another retrospective cohort study conducted in French with elevated AST (24.3%) and ALT (12.8%) levels but had elevated AST and ALT with higher elevated levels of GGT (25.3%) as compared with a study conducted in Egypt.^{40 44 45} The reason might be the differences in the percentage of patients with severe COVID-19 or differences in the percentage of study participants who consume alcohol. Due to excessive drinking the level of GGT increases. GGT is implicated in alcohol use by keeping intracellular glutathione, the body’s most abundant antioxidant, at adequate levels to protect cells from oxidative stress as a result of metabolism (eg, that of alcohol).⁴⁶

The current study found that patients who tested positive for SARS-CoV-2 had higher mean ALT values compared with those who tested negative. Patients with a positive test were more likely to have an elevated ALT level than the negative group (24.2 vs 18.7; p=0.0031). This result is in line with a retrospective study conducted at Columbia University’s New York-Presbyterian Hospital with ALT (28 U/L vs 21 U/L, p<0.001) values.⁴⁷ This may be due to the fact that hepatocyte cells in the liver express ACE2, which is the principal receptor for SARS-CoV-2.⁴⁸

The current study shows that the likelihood of having increased levels of AST and GGT were 3.0 times and 4.55 times higher, respectively, in SARS-CoV-2 when compared with COVID-19-free patients. Additionally, the levels of BILT and BILD were 7.9 and 3.72 times higher in SARS-CoV-2-positive individuals than in the control group, respectively. Similarly, different studies showed that the levels of LFTs, specifically AST, ALT and ALP, were increased.4048,50 This is possible because patients with COVID-19 may experience liver damage due to the virus directly attacking liver cells,⁵¹ targeting ACE2, a widely expressed protein on hepatocytes (2.6%), cholangiocytes (59.7%) and hepatic endothelial cells.⁵²

Most elevated levels of LFTs were not associated with alcohol consumption, except for ALP. Alcohol drinkers have a lower level of ALP or being an alcohol drinker lowers the level of ALP by 75% as compared with non-drinkers (AOR=0.25, 95% CI 0.08 to 0.78, p=0.017). This finding is supported by a study conducted at the Oak Ridge Institute for Science at Zhongnan Hospital of Wuhan University.^{53 54} However, this contradicts a study conducted at the Addis Ababa Millennium College COVID-19 care and treatment centre.⁴⁹ Alcohol consumers had significantly lower ALP activities compared with non-consumers. This is because consumption of alcohol reduces both innate and acquired immune activity, and it has been hypothesised that this habit is correlated with an increase in liver ACE2 receptors. In addition excessive alcohol consumption weakens the immune system, making it more susceptible to infection with SARS-CoV-2⁵⁵ and also lowers the level of ALP indicating that the duct epithelium is injured by SARS-CoV-2 itself.⁵⁴ For ALP, there was a positive relationship between alcohol intake up until about 35–40 years and lower ALP activity in advanced age (our study includes an advanced age group), and thereafter, higher alcohol intake was associated with slightly lower ALP.⁵³

The severe stage of COVID-19 was associated with hypoalbuminaemia, with 3.3 times lower odds than patients in mild and moderate stages (AOR=3.3, 95% CI 1.4 to 7.9). This is supported by an investigation done at Zhongnan Hospital of Wuhan University.⁵⁴ All individuals who are severely ill have hypoalbuminaemia, which is frequently related to inflammatory diseases. In this case, it is related to both a shortened albumin half-life (due to altered kinetics with neonatal Fc receptor downregulation and increased intracellular breakdown) and increased vascular permeability (due to augmented albumin distribution volume), resulting in a decrease in total albumin mass despite increased fractional synthesis. The majority of autopsy series have described COVID-19 as a capillary leak syndrome, which may considerably contribute to the occurrence of hypoalbuminaemia, which may then be interpreted as a marker of COVID-19 disease severity.⁵⁶

Pre-existing diabetes mellitus was associated with an elevated level of BILT and BILD. The likelihood that an elevated level of BILT and BILD reduces the risk of developing diabetic complication is 74% (AOR=0.26, 95% CI 0.09 to 0.77) and 72% (AOR=0.28, 95% CI 0.09 to 0.90), respectively. This result was inline with previous reports from India, China, and university of Gondar of Ethiopia.57,60 Chronic high glucose levels have toxic effects on the pancreatic islet’s structure and function. Chronic hyperglycaemia negatively impacts the beta cells in type 2 diabetes as well and results in chronic oxidative stress. As a reaction, our body produces more bilirubin as an endogenous antioxidant and anti-inflammatory.⁶¹ Additionally, increased levels of BILT may protect against the autoimmune, inflammatory-related pathology of type 1 diabetes and the oxidative physiological stress linked to the onset of type 2 diabetes.⁶² Additionally, physiological levels of BILT prevent the generation of various free radicals that can impair the cell’s ability to respond negatively to excessive glucose.⁶³ As a chain-breaking antioxidant, BILD, which binds weakly to albumin, is crucial in avoiding lipid peroxidation. The alternative possibility is that bilirubin directly inhibits inflammatory and immunological responses and acts as a signal of haem oxygenase-1 activity. BILD may have cytoprotective benefits, which are mediated by albumin-bound bilirubin preventing reactive oxygen species from damaging plasma and mitochondrial membranes at physiological levels.^{64 65}

The likelihood of elevated levels of BILT and BILD was 2.41 times and 3.7 times higher in men than women (AOR=2.41, 95% CI 1.2 to 4.9) and (AOR=3.7, 95% CI 1.72 to 8.2), respectively. This result was supported by previous studies done in the USA and India.^{66 67} Possible causes include the fact that oestrogens and testosterone have different effects on the conjugation rate and that uridine diphosphate-glucanosyltransferase is downregulated by testosterone while enzyme activity is increased by oestrogens.⁶⁸ Additionally, evidences from the animal model indicated that adult male rats had a higher activity of uridine diphosphate-glucuronosyltransferase as compared with adult female rats.⁶⁹

One of the limitation of the current study was the effect of antiviral drugs among COVID-19 cases were not analysed. The reason why we didn't compare between COVID-19 cases is that, all those in the COVID-19-infected group had taken antiviral drugs, so comparison was impossible. But analysing antiviral drugs as a predictor variable is importance to identify which antiviral drugs have an effect on LFT abnormalities (to compare the drug effect along with COVID-19).

Conclusion and recommendations

Conclusion

Patients with COVID-19 have decreased albumin levels, elevated AST, GGT, BILT and BILD levels, and more elevated laboratory test results were seen in severe stages. Factors such as alcohol consumption, diabetes mellitus and severe stage of the virus contribute to these abnormalities.

Recommendations

Patients with comorbidities should be followed up to manage LFT abnormalities, and they should also be careful about their diet. The nutritional condition of patients with COVID-19 should be evaluated as soon as they are admitted, and oral or intravenous nutrition support, either alone or in combination, should be taken into consideration (lower levels of albumin). Additionally, a follow-up investigation is necessary to manage abnormal LFTs caused by SARS-CoV-2. As the majority of the LFT abnormalities were seen in severely ill patients with COVID-19, physicians should provide special care for individuals who were in a severe stage of COVID-19, and patients with COVID-19 must also obtain health education on the likelihood of acquiring liver disorders and the most effective prevention against them. In addition, patients with comorbid diseases (diabetes mellitus) should be given follow-up care and medical education about the effects on LFTs.

The ministry of health should provide enough funding for facilities in order to monitor the LFTs of patients with COVID-19 infection routinely. It would be ideal to carry out further investigation on a large population and cohort study design, and the effect of alcohol consumption on the enzymatic activity of ALP by using tissue-specific forms or ALP iso-enzyme test should also be the preferable research area.

Strengths and limitations of the study

The comparison methodology we adopted is one of our study strengths. In this study, consistent definitions were used to reduce bias among readers. According to a search of the literature, this is the first study to assess the mean difference of LFTs and associated factors between SARS-CoV-2-positive individuals and healthy controls in the study area. The current study was also conducted in two different health facilities for better participant represenation and generalisation of results; and the majority of biochemical analytes used for the assessment of LFTs were done. Additionally, a laboratory test was performed at the referral hospital.

Despite the above strengths, our study has limitations. The study has a cross-sectional study design, and thus, it is not possible to show a cause-effect relationship with the use of convenient sampling techniques (lacks generalisability due to being vulnerable to selection bias). And also, in this study, laboratory screening test for selection of participants were not preformed to handle undiagnosed comorbidities. Although the effects of some viral drugs on LFTs were assessed, the effects of other medications used by study subjects on liver functions were not studied. Furthermore, a stored serum sample was used for the assessment of LFT abnormalities, which could alter the LFT results. An in-depth assessment of study subject recruitment was not done, which may result in recall bias by the study subjects.

Data availability statement

Data are available upon reasonable request.