The study was conducted at St. Paul’s Hospital Millennium Medical College (SPHMMC) which is located in Addis Ababa, Ethiopia, and was one of the main COVID-19 isolation and treatment centers in the country. It is a referral-specialized hospital, estimated to serve a total population of more than 5 million people. Patients from Addis Ababa city and rural areas around the city suspected of having COVID-19 were referred to the center. The COVID-19 isolation and treatment center had 260 beds delegated for COVID-19 patients. From 1 October 2020 to 16 September 2021, we used the national Ethiopian Public Health Guideline (EPHG) (19) and enrolled adult patients 18 years old and above who met the eligibility criteria for the study.

A time series cross-sectional study was conducted for 1 year from 1 October 2020 to 16 September 2021.

All adult patients (18 years old and above) having any acute respiratory illness (e.g., runny nose and sore throat) and at least one of the following symptoms: fever, cough, and shortness of breath were used as a source population of the study. All suspected COVID-19 patients who tested positive for SARS-CoV-2 by real-time PCR (RT-PCR) test and met the eligibility criteria were enrolled in the study. Our participants were all recruited from the local population and we did not encounter international patients that took the COVID-19 vaccine during the study period. Patients who were unwilling to participate and were previously known to have COVID-19 were excluded from the study. Once the study participants were enrolled in the study, they were followed up until discharge or death. All patients were isolated and stayed in the hospital until they tested negative for the SARS-CoV-2 by RT-PCR test. On average, mild/moderate COVID-19 patients stayed for 2 weeks, whereas those severe/critical patients stayed an average of 6 weeks and were followed up until discharge or death. Those participants who were discharged or died before the second set of measurements were taken were excluded from the study. COVID-19 patients enrolled in this study were diagnosed based on WHO interim guidance into mild (symptomatic patients without evidence of viral pneumonia or hypoxia), moderate (patients with signs of pneumonia and SpO2 ≥ 90% on room air), severe (signs of pneumonia plus one of the following: respiratory rate > 30 breaths/min; severe respiratory distress; or SpO2 < 90% on room air), and critical cases (patients that develop Acute Respiratory Distress Syndrome (ARDS), septic shock or multi-organ dysfunction) (20).

The study was conducted during the peak period of the first COVID-19 wave and our institute allocated a total of 260 beds for COVID-19 patient isolation and treatment. Thus, we found the convenient sample technique a more appropriate method for recruiting study participants. A consecutive convenient sampling technique was used to recruit study participants until the end of the study period. During the study period, a total of 1,512 COVID-19-suspected patients visited the SPHMMC COVID-19 isolation and treatment center. Of the suspected patients, 476 met the eligible criteria and 443 were willing to participate with a response rate of 93% (Figure 1). Oropharyngeal (OP) and/or nasopharyngeal (NP) samples were tested for SARS-CoV-2 RNA by RT-PCR and we excluded 164 patients who tested negative, and 279 patients tested positive of whom 11 participants were lost from the follow-up. Finally, we managed to include 268 study participants, and demographic, clinical, and laboratory parameters were recorded and assessed for potential predictors (Figure 1). Once the patients were admitted, blood samples were collected within 24 h of the time of admission, and a second blood sample was collected between 7 and 10 days after admission. COVID-19 patients were assessed for disease severity at admission and followed up till death or discharge from the hospital. We took samples proportionally throughout the year and the effect of a particular SARS-CoV-2 variant would not significantly affect the whole sample.

A pre-tested questionnaire and checklist were used to collect the study participants’ socio-demographic data, clinical data, and laboratory test results. The data were collected by trained physicians and laboratory technologists and reviewed by other investigators. Socio-demographic data were collected at the time of admission and clinical data were taken at the time of admission and during the patients’ hospital stay. Any information that could identify individual participants during or after data collection was kept private. Accordingly, patient clinical data on initial clinical presentation, comorbidities, and relevant treatment and clinical outcomes for all patients presented with acute respiratory symptoms or influenza-like illness symptoms were recorded. Oropharyngeal (OP) and/or nasopharyngeal (NP) swabs and a blood sample were collected within 24 h of patient admission and transported to SPHMMC laboratory for analysis and a second blood sample was collected between 7 and 10 days after admission for further assessment of patient progress. Collected data and measured parameters were pooled and cleaned at a global level for further analysis. A comparison of measured parameters was made between different groups like survivors vs. non-survivors and mild/moderate vs. severe/critical cases.

Oropharyngeal (OP) and/or nasopharyngeal (NP) swabs were collected according to the Ethiopian Public Health Institute Guideline (EPHI) (19). The samples were transported to the SPHMMC COVID-19 laboratory in VTM (viral transport media) (in a cold chain of 2-8°C) for RT-PCR analysis of SARS-CoV-2. Two hundred microliters (200 μL) of the NP or OP swab were mixed with 50 μL proteinase K and 200 μL lysis buffer that contains a guanidinium-based inactivating agent, and then viral RNA was extracted using a nucleic acid isolation Kit (Da’an Gene Corporation). Then, viral RNA was eluted with 60 μL elution buffer, and an RT-PCR reagent (Da’an Gene Corporation) was employed for SARS-CoV-2 detection. Two PCR primer and probe sets, which target the open reading frame 1ab (ORF 1ab/N) (FAM reporter) and nucleocapsid protein genes and VIC reporter genes, were added to the same reaction mixture. In each run, positive and negative controls were included. Samples were considered to be positive when both sets gave a reliable signal (≤40 CT value). We used primers and sequence-specific fluorescence probes designed and tailored to the highly conservative region of the COVID-19 genome. The important steps for RT-PCR amplifications took place at 50°C for 15 min, 95°C for 3 min, followed by 45 cycles at 95°C for 15 s and 60°C for 30s. All three cycles were accomplished within an hour and 35 min of the reaction being started.

The purpose of this antibody test is to confirm the participants’ immune response to the natural history of SARS-CoV-2 infection. The test was performed by the Elecsys anti-SARS-CoV-2 assay (Elecsys Anti-N; Roche Diagnostics, International Ltd., Rotkreuz, Switzerland) which can indicate the existence of anti-nucleocapsid antibodies during natural infection. The cut of index values below 1.0 (COI < 1.0) was interpreted as non-reactive or negative for anti-SARS-CoV-2 antibodies, while the cut of index values greater than or equal to 1.0 (COI ≥ 1.0) was interpreted as reactive or positive.

Fresh (<4 h from collection) whole blood samples (8–10 mL) were collected aseptically from each study subject using di-potassium EDTA–anti-coagulated vacutainer tubes (Becton Dickinson) within 24 h and 7–10 days after patient admission. The sample was analyzed using Beckman Coulter DxH 800 Hematology Analyzer (California, United States). The DxH 800 hematology analyzer measures 28 different hematological parameters like size, shape, and structure of the cell that can be recorded.

Serum/Plasma from coagulated blood was used for enzymatic and biochemical tests for liver and renal function assessment using Cobas C 501 Chemistry Analyzer (Roche) at the clinical chemistry laboratory. The instrument is fully automated and can analyze up to 60 assays like Glucose (Glu), Hemoglobin A1C (HBA1C), and Alanine Aminotransferase (ALT). The plasma sample was used for the quantitative determination of total cholesterol (CHOL), triglycerides (TG), high-density level cholesterol (HDL-C), and low-density level cholesterol (LDL-C), using Roche Cobas C311 (Roche/Hitachi).

The ion concentration was determined using the Roche-Cobas-C311 (Roche/Hitachi) clinical chemistry analyzer. The machine uses an ion-selective electrode (ISE) system to measure the serum’s sodium, potassium, and chloride levels. The ISE uses the unique properties of certain membrane materials to develop an electrical potential [electromotive force (EMF)] for the measurements of ions in the solution.

We used the Stago STA Compact Coagulation Analyzer (Diagnostica Stago), a robust and practical benchtop coagulation system to determine the different coagulation parameters. We measured prothrombin time (PT), a test that measures how long it takes for a clot to form in a blood sample, normally resulting in 10 to 14 s. The partial thromboplastin time [PTT; also known as activated partial thromboplastin time (aPTT)] was used to evaluate a person’s ability to appropriately form blood clots. It measures the number of seconds it takes for a clot to form in a sample of blood after reagents are added, normally resulting in 26–35 s, and an INR (international normalized ratio) is a type of calculation based on PT test results. In healthy people, an INR of 1.1 AU (Arbitrary Unit) or below is considered normal. An INR range of 2.0 AU to 3.0 AU is generally an effective therapeutic range for people taking warfarin for certain disorders.

All procedures and steps were pre-tested and proper amendments were taken before the actual data collection. For laboratory analysis, pre-analytical, analytical, and post-analytical stages of the laboratory work were done as per the Standard Operating Procedures (SOPs) of the lab to maximize the quality of collected data. Additionally, positive and negative controls were used in the laboratory analysis during each run. Quality controls were performed daily and after every additional calibration. We put effort into minimizing the missed data by simplifying the questionnaires and explaining the whole purpose of the study to increase the accuracy of the data.

The data were pooled together and entered into an Excel spreadsheet and subjects with incomplete information (those lost from follow-up) were removed from the analysis. The distribution of the data was checked using the Shapiro–Wilk test using R software (V 4.2.2, R Core Team, 2021), and means and Standard Deviation (SD) or median and interquartile ranges were used to describe contentious data with or without normal distribution, respectively, and the percentage was used for a categorical data set. A statistical test was done using SPSS version 23.0 (SPSS, Inc., Chicago, IL). For univariate analysis, Mann–Whitney U test, Chi-square, and Fisher’s Exact test were used to compare the two groups (alive vs. deceased as well as mild/moderate vs. severe/critical cases). The multicollinearity between the independent variables was checked using a Variance Inflation Factor (VIF), and a VIF of less than 3 was used for logistic regression. Independent variables with a p-value < 0.3 during univariate analysis were selected for the final logistic regression. The graphs were generated using the R 4.2.2 package (RcmdrPlugin.KMggplot2) (R Core Team, 2021). A 95% CI and p ≤ 0.05 was considered statistically significant.

The detailed study protocol was reviewed and clearance to carry out this study was obtained from the institutional review board (IRB) of St. Paul’s Hospital Millennium Medical College, Addis Ababa, Ethiopia. The objectives of the study were explained to all patients using the local language and patients were informed that they have the right not to participate or withdraw at any time and any stage from the study. The patients were also informed that their willingness to or not to participate does not affect their treatment at the hospital. The consent form was written in English and Amharic (the local language) so that the patient had full information about the study. For those illiterate patients, the full information and objectives of the study were explained verbally and written informed consent was taken from each participant. A unique code was given to participants and any information concerning the patients was kept confidential and the specimens collected from the patients were analyzed for the intended purpose only. Positive patient results were communicated timeously with the medical doctors who were working at a treatment center for better management.

A total of 443 adult patients showing signs and symptoms of COVID-19 (who met the eligibility criteria and were willing to participate) were isolated at St. Paul Hospital’s Millennium Medical College COVID-19 Center at the peak period of transmission from 1-Oct-2020 to 16-Sep-2021. After excluding 168 patients who tested negative for SARS-CoV-2 RNA by RT-PCR, 268 patients were enrolled for the final analysis (Table 1). The mortality rate among the participants was 21.6% (58/268). The median age of the participant was 56 years (IQR 44y – 66.25y) and the mean age of the participant was 55.5 years (SD = 15.3y, min 20y and max 98y). Of the total participants, women comprised 37.3% (n = 100) and men accounted for 62.7% (n = 168). Participants from the urban area were 82.2% (n = 222) and those that came from the rural part of the city were 17.8% (n = 46). Of the different symptoms of COVID-19, most participants presented with cough (91%; n = 244), short breath (92.2%; n = 247), and loss of appetite (85.4%; n = 229). We found that the age of the participant at the time of admission significantly (p < 0.05) affected the final outcome of the COVID-19 patients. Those COVID-19 patients who died had a significantly higher (p < 0.05) median age of 61 years (50 y – 70 y) at the time of admission compared to those who survived and had a median age of 54 years (IQR 42y – 65y) (Table 1), which align with another study (21). On the other hand, we found no significant association (p > 0.05) with gender, presence of comorbidity, marital status, education status, and clinical symptoms at the time of admission on the final outcome of the participants (Table 1). There was also no significant association (p > 0.05) between a particular comorbidity with the final outcome of COVID-19 patients (Supplementary Table S1). However, we observed that even if some of the COVID-19 patients who died were free of any disease, they were relatively old [X, IQR (60 (56–70))] compared to those who survived and were free of the disease [X, IQR (51 (38–60))] (p < 0.05) (Supplementary Table S2) which could explain our loss of association between comorbidity with fatal outcome of COVID-19, and age was one of the risk factors of COVID-19’s fatal outcome.

We measured vital clinical signs at the time of admission, and we found that the level of oxygen saturation (Figure 2A), systolic blood pressure (Figure 2B), and diastolic blood pressure (Figure 2C) were significantly (p < 0.05) lower in deceased individuals compared to survivors of COVID-19. On the other hand, body temperature, pulse rate, respiration rate, and random glucose (Figures 2D–G) did not show any significant (p > 0.05) difference between the two groups (22). We then looked into the different hematological and clinical parameters measured at the time of admission and assessed their role in determining the outcome of COVID-19 after admission. Univariant analysis of our data showed those COVID-19 patients who survived had significantly (p < 0.05) higher lymphocyte and monocyte percentages (Table 2). Others have also shown that COVID-19 patients with low lymphocyte counts have a high mortality rate (15), whereas the percentage of neutrophil and neutrophil-to-lymphocyte ratio (NLR) were significantly lower (p < 0.05) in COVID-19 patients who survived compared to those who died (Table 2). The increased NLR level has been linked with COVID-19 disease severity and mortality (23). The RBC count of COVID-19 patients was higher (p < 0.05) in patients who survived 4.9 × 10⁶/μL (IQR, 4.3 × 10⁶/μL – 5.2 × 10⁶/μL) compared to those who died 4.4 × 10⁶/μL (IQR, 3.82 × 10⁶/μL – 5.02 × 10⁶/μL) (Table 2). As the normal count of RBC was lower in women (4.2–5.4 × 10⁶/μl) compared to men (4.7–6.1 × 10⁶/ul), we assessed whether the effect of RBC count in COVID-19 patients who survived was still higher than in deceased patients independent of gender. We separately analyzed the data for the RBC level between the two genders and we observed a similar trend, that is, a lower RBC count in deceased COVID-19 patients in both male and female patients (Supplementary Figure S1). Similarly, at the time of admission, the liver function test ALP was significantly lower (p < 0.05) in survivors [X (IQR), 70 IU/L (57.25 IU/L–86.75 IU/L)] compared to deceased [X (IQR), 88 IU/L (66.5 IU/L–109.5 IU/L)] patients. In the same manner, at the time of admission, the liver function test GOT, the kidney function test creatinine and urea, and the level of K concentration were lower (p < 0.05) in survivors compared to deceased patients (Table 2). A multivariate analysis of measured parameters that have a p-value < 0.3 was analyzed, and neutrophil percentage and RBC count were important predictors of COVID-19 outcome (Table 2).

We categorized the COVID-19 patients into two groups based on their clinical status at the time of admission and assessed the measured parameters. Of the total COVID-19 patients included in the study, 203 (75.7%) presented with severe/critical COVID-19 and 65 (24.3%) patients manifested mild/moderate COVID-19 symptoms. Depending on the COVID-19 severity, patients were followed up from 2 weeks to 6 weeks. A significant association (p < 0.05) was found between the clinical status at the time of admission with the final outcome of the disease, where most deaths (89.6%; n = 52) occurred in severe/critical cases compared to 10.4% (n = 6) in mild/moderate COVID-19 patients (Figure 3A) which align with another study (24). The clinical status of the participants at the time of admission was not associated (p > 0.05) with gender (Figure 3B) and no significant difference (p > 0.05) in age between mild/moderate [X, IQR (56y, 40y–67.75y)] compared to severe/critical cases [X, IQR (58y, 42.5y–68.2y)] was observed at the time of admission (Figure 3C). Of the different measured parameters, lymphocyte percentage was significantly higher in mild/moderate [X, IQR (12, 7.5–12%)] compared to severe/critical [X, IQR (9, 5.25–18%)] patients at the time of admission (Figure 3D). On the other hand, other measured parameters like the percentage of neutrophils, the percentage of monocytes, and RBC count did not show a significant difference (p > 0.05) between the two groups of patients at the time of admission (Figures 3E–G). We also analyzed the different laboratory parameters in the presence and absence of comorbidity and we found that the presence of comorbidity has no significant (p > 0.05) impact on these laboratory parameters and in the final outcome of COVID-19 patients (Supplementary Figure S2).

We further looked into the natural antibody production of 201 participants and assessed its association with the final outcome and the clinical status of COVID-19 patients. We observed that severe/critical patients produce a much higher level of natural antibodies (p < 0.05) compared to mild/moderate COVID-19 patients (Figure 4A), whereas the outcome of COVID-19 patients was not associated with the production of natural antibodies against SARS-CoV-2 (Figure 4B).

We then measured different hematological and clinical parameters after the patients were hospitalized (within 7–10 days) and assessed their association with the patient’s final outcome. COVID-19 patients who survived showed significantly (p < 0.05) higher lymphocyte percentage (Figure 5A), monocyte percentage (Figure 5D), and RBC count (Figure 5E) after hospitalization compared to COVID-19 patients who died. On the other hand, COVID-19 patients who survived showed decreased (p < 0.05) neutrophil percentage (Figure 5B) and neutrophil-to-lymphocyte ratio (Figure 5C), compared to those deceased COVID-19 patients. Likewise, the kidney function test Urea (Figure 5F), creatinine (Figure 5G), and liver function test ALP (Figure 5H) were significantly lower after hospitalization in COVID-19 patients who survived compared to those who died.

We finally looked into patients who were presented as critical/severe cases at the time of admission, and those who were able to survive showed different measured parameters compared to critical/severe patients who died. We observed that critical/severe patients who survived had significantly (p < 0.05) increased lymphocyte (%) (Figure 6A), monocyte (%) (Figure 6D), and RBC count (Figure 6E) and significantly (p < 0.05) decreased neutrophil (%) (Figure 6B), NLR (Figure 6C), and creatinine, urea, and ALP levels (Figures 6F–H) compared to critical/severe patients who died. This could suggest that favorable outcomes can be achieved by monitoring critically ill COVID-19 patients to maintain measured parameters.