Predictors of Mortality Among Hospitalized COVID-19 Patients at a Tertiary Care Hospital in Ethiopia

Abstract

Background

The very unprecedented virus causing severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has continued causing catastrophes in economy and loss of human lives. Despite countries’ urgent and resilient public health actions against the COVID-19 pandemic, the disease is causing a large number of deaths. However, predictors of mortality among hospitalized COVID-19 patients have not been well investigated in Ethiopia. Therefore, this study aimed to identify the predictors of mortality among hospitalized COVID-19 patients at a tertiary care hospital in Ethiopia.

Methods

A hospital-based retrospective cohort design study was implemented among hospitalized COVID-19 patients at a tertiary care hospital in Harar, Ethiopia from March 20 to August 20, 2021. Data of 531 admitted patients were entered using Epi-data 3.1 and exported to STATA 14.2 for analysis. Binary logistic regression was used to identify significant predictors of outcome variables with an adjusted odds ratio (AOR) with a 95% confidence interval.

Results

Of the total 531 study participants, 101 deaths occurred. The mortality rate was 16.2 per 1000 person-days of observation with median survival time of 44 days with IQR [28, 74]. Smoking history [AOR=2.55, 95% CI (1.15, 5.65)], alcohol history [AOR=2.3, 95% CI (1.06, 4.97)], comorbidities [AOR=2.95, 95% CI (1.26, 6.91)], and increasing oxygen saturation [AOR=0.92, 95% CI (0.89, 0.95)], and lymphocyte count [AOR=0.90, 95% CI (0.88, 0.97)] were independent significant predictors of death from Covid-19.

Conclusion

The incidence of mortality among hospitalized COVID-19 patients was found to be high. Devising individual, tailored management for patients with “risk” behaviors, comorbid conditions, and poor prognostic markers such as lymphopenia and low oxygen saturation, may reduce the incidence of death among hospitalized COVID-19 patients.

Introduction

The very unprecedented virus causing severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) which causes coronavirus disease 2019 (COVID-19) was detected in Wuhan, China in December 2019.¹ The virus spread throughout the world and was categorized as a pandemic by the World Health Organization (WHO) in March 2020.² Coronavirus causes respiratory diseases and most commonly spreads from individual-to-individual through infected air precipitation, people making contact with fingers or faces that contain the virus and then touching their eyes, nose, or mouth with unclean hands.²

As of November 08, 2021, COVID-19 had infected more than 250,742,353 people globally, of whom about 5,067,409 had died, and 226,971,946 had recovered from COVID-19. In Ethiopia, 367,210 people have been infected by the virus of whom 6542 died, giving overall mortality rate of 18 deaths per 1000 infected persons.³

Adding to the current COVID-19 pandemic, based on the recent epidemiological update by the WHO, as of June 22, 2021, four SARS-CoV-2 variants of concern (VOCs) have been identified since the beginning of the pandemic. Alpha: first variant of concern described in the United Kingdom (UK) in late December 2020, Beta: first reported in South Africa in December 2020, Gamma: first reported in Brazil in early January 2021, Delta: first reported in India in December 2020.^4,5 This VOC has the potential to spread faster with higher mortality than first seen in the pandemic to date.⁶

Studies revealed that the outcome of COVID-19 varies and is mainly determined by individual characteristics. The outcome of COVID-19 infection could include spontaneous recovery, respiratory complications that require mechanical ventilation, sepsis and septic shock, and multi-organ failure, including acute kidney injury and cardiac injury and death.^7–10 Clinical features like fever, shortness of breath, deranged laboratory findings have been identified as indicators of poor prognosis.^11,12

Older age and underlying comorbidities such as hypertension, cardiovascular disease and diabetes, have been reported as risk factors for severe disease and death.^10,13 To mitigate some of the risks in patients with chronic disease, technological innovation such as telemedicine along with public health strategies were considered in some countries.¹⁴ Telemedicine offsets the decline in outpatient visits during COVID-19 while providing critical patient continuity and limiting exposure to health systems and healthcare workers.¹⁵

Non-pharmacological measures such as lockdown and travel restrictions have shown a positive result in terms of containing the transmission of the virus. Studies revealed countries that implemented an immediate application of lockdown reduced deaths compared to countries that delayed the application of this strong containment measure.^16,17 Vaccines have been found to be highly efficacious at preventing symptomatic disease, as shown by clinical trials^18,19 and real-world evidence.^20–22 As of November 05, 2021 in Ethiopia, 5,237,750 doses of COVID-19 vaccines had been administered.²³

Literature has indicated that COVID-19-related morbidities and mortalities in sub-Saharan African countries are less as compared to higher income countries.^24,25 Different assumptions for this have been forwarded, including the higher proportion of younger population in SSA, under-reporting of mortality, low coverage of COVID-19 testing, different immune co-activation status and exposure to non-pathogenic corona viruses.^26–29

Having knowledge of the incidence and predictors of death from COVID-19 infection in our setup is the key to make an informed decision to obtain the most favorable outcome of treatment for each patient. Previously conducted studies in Ethiopia are not adequate in addressing important clinical predictors of severity and mortalities from COVID-19 infection.^30,31 Therefore, the aim of the study was to investigate the factors associated with mortality in hospitalized COVID-19 patients at a tertiary care hospital in Ethiopia.

Methods and Materials

Study Area, Period, and Design

A hospital-based retrospective cohort study was conducted in Harari Regional State, Ethiopia from March, 2020 to August, 2021. The region is one of the ten regional administrations in Ethiopia and located 522 kilometers from Addis Ababa to the east of the country. Hiwot Fana Specialized University Hospital is one of the largest referral and teaching hospitals in the region serving the large majority of Eastern Ethiopia estimated to be more than six million population in the catchment areas. The hospital is one of the ten regional centers designated by the Ethiopia Ministry of Health to treat COVID-19 cases in Eastern Ethiopia.

Population and Sample

The source population of the study was all COVID-19 patients admitted to the hospital. The study population was all COVID-19 hospitalized patients, between March, 2020 and 20 August, 2021. All COVID-19 patients admitted during the study period were included in the study using a retrospective approache. Those patients who had incomplete data for variable of interest were excluded from the study.

The sample size was determined using double population proportion under the assumptions of type-I error=0.05, power (1-β=0.8), proportion of survival group= 0.57, and proportion of deceased patients= 0.43.³² The sample size was determined to be 400. Finally, by considering 10% of loss of follow-up, the sample was determined to be 440. However, to increase the precision of the result we included all 531 COVID-19 patients whose treatment outcome was identified before 20 August 2021.

Study Variables and Measurements

The outcome variable was survival status of COVID-19 cases (discharged vs died). The independent variables included socio-demographic characteristics such as gender, age, marital status, occupation, residence, COVID-19 clinical characteristics, comorbidities, and clinical outcomes, use of ventilators, COVID-19 clinical management, laboratory findings, and imaging findings. Diagnosis of COVID-19 was defined as the patient having a positive result on the oropharyngeal swab for SARS-CoV-2 by reverse transcriptase polymerase chain reaction (RT-PCR). Common radiological assessment included chest radiograph and computed tomography of the chest based on clinical decision-making. COVID-19 clinical stages are defined as follows.³³

Asymptomatic infection: individuals who test positive for SARS-CoV-2 using a virology test (ie, a nucleic acid amplification test [NAAT] or an antigen test) but who have no symptoms that are consistent with COVID-19.

Mild illness: individuals who have any of the various signs and symptoms of COVID-19 (eg, fever, cough, sore throat, malaise, headache, muscle pain, nausea, vomiting, diarrhea, loss of taste and smell) but who do not have shortness of breath, dyspnea, or abnormal chest imaging.

Moderate illness: individuals who show evidence of lower respiratory disease during clinical assessment or imaging and who have oxygen saturation (SpO₂) ≥94% on room air at sea level.

Severe illness: individuals with clinical signs of pneumonia (fever, cough, dyspnea, fast breathing) plus SpO₂ <94% on room air, respiratory rate >30 breaths/min, or lung infiltrates >50%

Critical illness: individuals who have respiratory failure, septic shock, and/or multiple organ dysfunctions.

Equipment Used and Preparation Made on the SARS-COV-2 Assay Using RT-PCR

Viral transport medium (VTM), swab sticks, ice-box, ice-pads, tongue depressor, marker, requisition form, and personal protective equipment (PPE) were prepared and used beforehand. The VTM that contains three mL fluids composed of gelatin and antimicrobial agents in a buffered salt solution was used. This VTM was used to prevent the sample collected from drying, maintains the viability of the virus, and avoids the growth of contaminants. The used swabs are made up of rayon with a plastic shaft. Ice-box and ice-pads are used for maintaining a cold chain during sample transportation from the sample collection area to the laboratory center.³⁴ The ice-pad was filled with water and stored in the freezer (−20°C) before and after use.

Sample contamination of the nasal vestibule was avoided by sterile openening of the outer case of the swab, inserting the swab into the mouth by slightly elevating the tip of the nose, and keeping the tip of the swab in the oropharynx for a few seconds, then rotating to achieve the highest absorption of oropharyngeal secretions.^35,36

Data Collection Procedure and Quality Control

A checklist was prepared and used to collect relevant data by reviewing patient charts. Health professionals were assigned as data collectors and data were extracted by reviewing follow-up charts and cards of patients. Two-day training was given to the data collectors on the objective of the study and how to review the documents as per the data extraction checklist. The principal investigator supervised the data collectors closely. Adequacy of the checklist was evaluated and ambiguous questions were modified before the actual data collection. Besides, daily monitoring of data for completeness and consistency was done.

Statistical Analysis

Data were entered using Epi-data 3.1 and exported to STATA 14.2 for analysis. Descriptive statistics were conducted using frequency and percentages for categorical measurements and mean and median were used to summarize continuous variables.

The important assumptions of Chi-square were checked before statistical modeling where chi-square assumptions were; sample was randomly drawn from the population and all expected values were are at least 5. In addition, important assumptions such as multicollinearity, independence of errors, and linearity in the Logit for continuous variables of logistic regression were checked on top of modeling. Moreover, a receiver operating characteristic curve (ROC-curve) was used to assess the trade-off between sensitivity and specificity. For each predictor, bi-variable binary logistic model was fitted. In bi-variable analysis, predictors with a p-value less than 0.2 were included in multivariable model. Finally, multivariable logistics regression model was fitted for clinical outcome of COVID-19 patients (discharged and death) with covariates fulfilling the prerequisite. The association between outcome and predictors was presented by adjusted odds ratios (AORs) with 95% confidence interval. The statistically significant association between outcome variable and predictors was detected by using 95% confidence interval. The confidence interval which does not include the null hypothesis value means there is statistically significant association. Moreover, after fitting the model, goodness of the final model was checked by using the Hosmer-Lemeshow test. The Hosmer-Lemeshow statistic indicates a good fit at p-value of 0.05 or greater.³⁷ If the confidence interval does not contain the null hypothesis value, the results are statistically significant.

Ethical Statement

Haramaya University College of Health and Medical Sciences Institutional Health Research Ethical Review Committee (IHRERC) ethically cleared the paper. A letter of permission was sent from the College of Health to Hiwot Fana specialized University Hospital. Due to difficulty in reaching patients to get informed consent, data were obtained from patients’ medical records anonymously guaranteeing information confidentiality, and all data extraction processes were conducted per the declaration of Helsinki.

Results

Socio-Demographic Characteristics

Data of 531 patients who were at COVID-19 treatment center between May 2020 and August 2021 were collected. The final analysis was done on the data of 531 patients. The median age of study subjects was 40 years with an interquartile range (IQR) of 28–60 years (Table 1).

Table 1.

Baseline Socio-Demographic Characteristics of Covid-19 Patients in Hiwot Fana Specialized University Hospital Treatment Center, Harar, Ethiopia

Variables	Frequency (%)
Sex
Male	342 (64.41)
Female	189 (35.59)
Age group
1–25	83 (19.67)
26 −50	218 (51.66)
51–65	64 (15.17)
>65	57 (13.51)
Ethnicity
Harari	294 (55.37)
Oromo	224 (42.8)
Other	13 (2.44)
Marital status
Single	79 (14.88)
Married	393 (74.01)
Divorced	4 (0.75)
Widowed	11 (2.07)
Separated	44 (8.29)
Occupational status
Employed	82 (17.40)
Unemployed	154 (32.70)

Note: Other: Somali, Amhara.

Clinical and Baseline Behavioral Characteristics

The median respiratory rate of the patients was 8 BR with IQR: (24, 40). The median heart rate of the study participants was 100bit/m with [IQR: 219 to 464]. In the baseline of all, 116 (21.85%) of study subjects were alcohol consumers (Table 2). One hundred and one (19.06%) study participants were in the critical COVID-19 clinical stage and 84 (15.85%) were in severe COVID-19 clinical stage (Table 3).

Table 2.

The Median with Interquartile Range of Clinical Biomarker Characteristics of Covid-19 Patients in Hiwot Fana Specialized University Hospital Treatment Center, Harar, Ethiopia

Variables	Median	IQR
Partial oxygen saturation	89	[76, 95]
Alanine transferase	39	[76, 95]
Aspartate	45	[32, 65]
Lactate dehydrogenase	45	[36, 150]
Bilirubin	0.6	[0.60, 0.90]
Creatinine	0.9	[0.70, 1.30]
Bun level	45	[28, 62]
Neutrophil	78.70	[62.70, 88.82]
Lymphocyte	15.30	[8.10, 27.00]
Leucocyte	8.46	[5.67, 13.07]
Hemoglobin	12.9	[10.9, 14.76]
Platelets level	276.5	[201.5, 352, 25]

Table 3.

Frequencies and Percentages of Clinical and Baseline Behavioral Characteristics of Covid-19 Patients in Hiwot Fana Specialized University Hospital Treatment Center, Harar, Ethiopia

Variables	Frequency (%)
Baseline smoking
Yes	191 (35.97)
No	340 (64.03)
COVID-19 clinical stage
Mild	154 (29.06)
Moderate	191 (35.96)
Severe	85 (16.00)
Critical	101 (19.02)
Medication history
Yes	217 (40.87)
No	314 (59.13)
Cough
Yes	417 (21.47)
No	114 (78.53)
Fatigue
Yes	160 (30.13)
No	371 (89.87)
Fever
Yes	173 (32.58)
No	358 (67.42)
Diarrhea
Yes	33 (6.21)
No	498 (93.79)
Asymptomatic
Yes	70 (13.18)
No	461 (86.82)
X-ray finding
Bilateral consolidation	110 (20.72)
Bronchopneumonia	22 (4014)
Opacity	22 (4014)
Effusion	8 (1.51)
Normal	354 (66.67)
Other	6 (1.13)
Comorbidity
Yes	280 (41.27)
No	251 (52.73)
Dexamethasone
Yes	96 (18.08)
No	435 (81.92)
Mechanical ventilation
Yes	339 (63.84)
No	192 (36.14)
Immediate cause of death
Multi-organ failure	52.47
Acute respiratory distress syndrome	47.53
Shortness of breath
Yes	186 (35.03)
No	345 (64.97)

Incidence of Mortality

The patients were followed for a minimum of 1 and a maximum of 74 days with a median follow-up time of 8 days [IQR, (5, 14)]. During the follow-up time of 531 study participants, 101 patients died. The study demonstrated that the incidence of mortality was 16.2 per 1000 person-days of observation with median survival time of 44 days with IQR [28, 74].

Model Assumption and ROC Curve

Multicollinearity assumption was checked by using variance inflation factor (VIF) where variance inflation factor (VIF) is a measure of the amount of multicollinearity in a set of multiple regression variables. For this study the VIF of predictors was less than 3 so we can conclude that there was no multicollinearity (Table 4). The Tietjen-Moore test was used to detect strongly influential outliers in the data set. Moreover, linearity in the Logit assumption was checked using the Box-Tidwell variables transformation. Finally, ROC curve was used to see the trade-off between sensitivity and specificity. The result from ROC curve showed that the curve far from the curve 45-degree diagonal of the ROC space in the right side. So we are confident that the predicted probabilities from the model can accurately take on all possible values between 0 and 1 or it is (Figure 1).

Table 4.

Variance Inflation Factor (VIF) of Predictors Associated with Covid-19 for Patients at Treatment Center in Hiwot Fana Specialized University Hospital Treatment Center, Harar, Ethiopia

Variables	Variance Inflation Factor (VIF)
Occupational status	1.23
Alcohol utilization	1.45
Medication history	2.10
Cough	1.55
Fever	1.06
Comorbidity	1.15
Age category	2.03
Respiratory rate	1.82
Heart rate	1.04
Partial Oxygen saturation	1.26
Creatinine	2.01
Lymphocyte	1.76
Dexamethasone	1.45

Figure 1.

Receiver operating characteristic curve for death among Covid-19 patients at a tertiary care hospital in Harar, Ethiopia.

Predictors of Mortality

From the final multi-variable binary logistic regression model: smoking status, alcohol consumption, having comorbidity, oxygen saturation, and lymphocyte count had independent significant association with death from COVID-19 (P<0.05).

The risk of death among COVID-19 patients who had smoking history was 2.55 [AOR=2.55 CI (1.15, 5.65)] times higher when compared with non-smokers. In addition, of “risk” behavior factors, risk of dying was 2.3 [AOR=2.3, CI (1.06, 4.97)] times more likely among COVID-19 individuals who had history of alcohol consumption when compared to their counterparts. The risk of death among COVID-19 patients who had comorbidities was 0.95[AOR=2.95, CI (1.26, 6.91)] times more likely when compared with patients who had no comorbidity. The current study indicated that increment in oxygen saturation level to optimum stage decreases risk of death from COVID-19 by 8% [AOR=0.92 CI (0.89, 0. 95)]. In addition, the current study revealed that one unit increment in lymphocytes among COVID-19 hospitalized patients decreases risk of death by 10%(AOR=0.90, 0.88–0. 97)] (Table 5).

Table 5.

Predictors Associated with Covid-19 for Patients at Treatment Center in Hiwot Fana Specialized University Hospital Treatment Center, Harar, Ethiopia

Variables	Covid-19 Outcome		COR With 95% CI	AOR With 95% CI	P-value
	Death	Discharged
Occupational status
Employed	75	206	1	1
Unemployed	26	224	0.31[0.19, 0.51]	1.22 [0.48,3.04]	0.21
Smoking Status
Yes	63	127	4.12[2.62, 6.50]	2.55 [1.15, 5.65]*	0.02
No	37	303	1	1
Alcohol utilization
Yes	42	74	3.42[2.14, 5.46]	2.30 [1.06,4.97]*	0.03
No	59	356	1	1
Medication history
Yes	77	140	6.64[4.02, 10.96]	1.25[0.52, 3.01]	0.26
No	24	290	1	1
Cough
Yes	88	329	2.07[1.11, 3.87]	1.08[0.36, 3.22]	0.34
No	33	101	1	1
Fever
Yes	45	117	3.33[2.13, 5.20]	1.40[0.66, 2.95]	0.22
No	56	313	1	1
Comorbidity
Yes	94	186	17[7.98, 38.86]	2.96[1.26, 6.92]*	0.01
No	26	206	1	1
Dexamethasone
Yes	34	62	3.02[1.80, 4.92]	1.63 [0.74, 3.57]	0.44
No	67	368	1	1
Age category (in years)
1–25	96	9	1	1
25–55	229	44	1.56[0.70, 3.55]	0.44[0.16, 2.58]	0.36
56–65	56	24	3.98[1.85, 859]	0.57[0.18, 3.34]	0.53
>65	49	24	5.22[2.25, 12.09]	0.93[0.14, 5.96]	0.37
Respiratory rate	–	–	1.06[1.04, 1.08]	1.03[0.99, 1.07]	0.13
Heart rate	–	–	1.03[1.02,1.04]	1.02[0.98, 1.03]	0.47
Partial Oxygen saturation	–	–	0.90[0.88, 0.92]	0.93[0.88, 0.96]*	0.00
Creatinine	–	–	1.56[1.22, 2.00]	1.09[0.88, 1.34]	0.41
Lymphocyte	–	–	0.90[0.88, 0.94]	0.92[0.88, 0.97]*	0.00

Note: *Statistically significant variables (p< 0.05).

Discussion

The study demonstrated that the incidence of mortality was 16.2 per 1000 person-days of observation with median survival time of 44 days with IQR [28, 74]. Smoking status, alcohol history, comorbidity, oxygen saturation, and lymphocytes had significant association with mortality.

In this study, the proportion of mortality was 19.2%. This is comparable with the study finding from Italy (16.7%).³⁸ However, the present finding is inconsistent with studies from Peru,³⁹ France,⁴⁰ and reports from multiple other countries.⁴¹ Pradhan and Olsson confirmed that deaths from COVID-19 are varied across the world.⁴² This death burden variation could be due to socio-demographic differences such as patients’ age, gender, living environment, personal behaviors, pre-COVID-19 medical conditions, and setting in which the patients received the care.^41,43–45

COVID-19 patients who had smoking history had higher risk of death (2.55 times) than non-smoking COVID-19 patients. This finding is in line with several studies conducted in different parts of the world.^46–52 This could be due to the fact that cigarette smoking increases ACE2 expression at the apical surface of the airway epithelium that results in the exacerbation of viral entry that imposes poor clinical progression of people with SARS-CoV-2.^53,54 In addition, cigarette smoking is associated with poor lung function that could worsen in COVID 19 infections and finally result in death. In addition, both smoking cigarettes^55–57 and SARS-COV-2^58–60 affect lung functions. Thus, the overlapping effect of the two conditions could exacerbate lung impairment that in turn results in poor prognosis of COVID-19 smoker patients.

Even if gender was not significant in the final model, the majority of deaths occurred among males compared to females. The studies suggest that more males are dying from COVID-19.^42,61,62 This could be due to variation in immune system responses that may contribute to viral load, clinical stages and outcomes, and hormonal effects.^42,63,64 Moreover, males had higher probability of requiring intensive care treatment,⁶⁵ due to developing severe conditions,⁶⁶ than females.

COVID-19 patients who had alcohol history had 2.3 times higher risk of dying compared to non-alcohol consuming COVID-19 patients. This finding is aligned with studies conducted in India,⁶⁷ United States,⁵² and United Kingdom.^68,69 This could be explained by the fact that alcohol consumption intensifies organ sicknesses, especially liver and heart, and leads to increased risk of severity of disease^70–72 and could also independently affect the clinical outcome of the patients.

COVID-19 patients who had comorbidities had 2.95 times higher risk of death than patients who had no comorbidity on presentation. This finding is in line with findings from⁷³ Denmark,⁷⁴ Romania,⁷⁵ Italy,^76,77 Mexico,⁷⁸ and Michigan state, USA.⁷⁹ This could be due to the fact that co-existence of SARS-COV-2 infection with other disease could have immunosuppressive effect and low treatment response that might negatively affect the treatment outcome of COVID-19 patients.^80,81 Fatality could be exacerbated due to cytokine storm, an acute hyperinflamatory response formation accounted for by comorbidities among COVID-19 patients.⁸²

Oxygen saturation increment by one unit decreases risk of death by 8% among COVID-19 patients. This finding is supported by study findings from Italy,^76,83,84 Peru,^39,85 Mexico,⁸⁶ Iran and Afghanistan.⁸⁷ It is the fact that low oxygen saturation, hypoxia leads to progression of COVID-19 critical stages,⁸⁸ thus, low level of oxygen saturation could result in early death of COVID-19 patients from the alveoli damage that impairs oxygen exchange with carbon dioxide.⁸⁹

Furthermore, lymphocyte increment by one unit decreases the risk of death by 10% among COVID-19 hospitalized patients. This finding is supported by studies conducted in Italy,⁷⁶ Mexico,⁹⁰ Spain,⁹¹ France,⁴⁰ and Wuhan.⁹² Studies showed that as COVID-19 clinical stages advance, the lymphocyte count decreases,^93–95 thus, as the clinical stages of COVID-19 patients advance to severe and critical, the clinical outcome could be bad. This study may serve as a basis for similar setups because there has been no prior information such as this from similar setups. Even if this study has paramount strengths, it would not be free from limitations. These limitations could be missing some important factors associated with COVID-19-related deaths such as ambient temperature, atmospheric pollution, social distancing, and obesity due to nature of secondary data. Another limitation could be that this study was conducted in only a single center.

Conclusion

This study demonstrated that the incidence of mortality was 16.2 per 1000 person-days of observation with median survival time of 44 days. The predictors of mortality among COVID-19 patients were current smoking, alcohol drinking, having comorbidity conditions; low oxygen saturation, and low lymphocyte count. Therefore, patients with risk behaviors, comorbidities, lymphopenia, and hypoxia need close monitoring and management during health care service in clinical setups.

Data Sharing Statement

The supporting data of this study are available on reasonable request.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no conflicts of interest for this work.