Coronaviruses are a large family of viruses that can cause a variety of illnesses in humans, ranging from the common cold to Middle East Respiratory Syndrome (MERS) and Severe Acute Respiratory Syndrome (SARS) [1] . The novel human coronavirus disease COVID-19 has become the fifth documented pandemic since the 1918 influenza pandemic [2] . This new virus is the causative agent of this new respiratory infectious disease called COVID-19 (Coronavirus Disease 2019) [3] . This new virus is associated with a mortality rate of 0.1 to 1%, likely an overestimate given uncertainties regarding the denominator. Among identified patients, 15% of cases will develop a severe form, and mortality in intensive care units affects 30 to 70% of cases [4] . In Gabon, the first case was identified on March 12, 2020 [5] , and the number of cases increased rapidly and in waves, with a first peak in June 2020, then a second in April 2021, and finally a third peak in January 2022. The clinical spectrum of the disease is broad, ranging from minimally symptomatic forms to extremely severe forms with life-threatening respiratory distress syndrome, requiring intensive care. The literature has increasingly confirmed these various points on the prognostic factors of patients with SARS-CoV-2, however in Gabon and the sub-region, these were mainly general patients [6] [7] . Studies specifically focusing on intensive care patients are rare.

The intensive care unit of the Omar BONGO ONDIMBA Army Training Hospital (OBO ATH) admitted its first patient on December 4, 2020. Since then, during this pandemic, it has admitted a growing number of patients with diverse profiles and sometimes discordant developments, in a pathology where the severity factors were quickly individualized, particularly advanced age [7] [8] . In fact, in a relatively young Gabonese population, it seemed useful to us to focus particularly on severe COVID patients, in their profile and evolution. We therefore carried out a study whose objective was to describe the epidemiological characteristics of the patients, and the therapies deemed necessary, in the intensive care unit of the Omar BONGO ONDIMBA Army Training Hospital.

The study took place at the Anesthesia-Resuscitation and Burns Department of the Omar BONGO ONDIMBA Military Training Hospital. This was a descriptive, retrospective, single-center study of patients with SARS-CoV-2 pneumonia. Our study took place over a two-year period, from December 2020 to December 2022. The study population consisted of any patient admitted to the intensive care unit for respiratory distress during the study period.

Our study included all patients admitted to the Anesthesia and Intensive Care Unit of the OBO Hospital during the study period for SARS-CoV-2 infection. The diagnosis was made based on the following clinical and paraclinical findings:

Clinically:

On the paraclinical level:

The following were not included:

Data collection was conducted by reviewing patient records on a survey form including:

Data were entered and processed using Excel© 2019 software. Analysis was performed by describing the sample studied according to sociodemographic, epidemiological, clinical, paraclinical, and therapeutic characteristics. Categorical variables were expressed as percentages, and quantitative variables as mean and standard deviation, as well as median and interquartile range. The Chi-square test was used to verify the association between categorical variables, and the student t-test was used to compare means. The significance threshold was set at p ≤ 0.05.

Authorization was obtained from the Chief Physician in charge of the OBOATH and the Head of the Anesthesia-Resuscitation Department for access to patient records. Data confidentiality was respected.

A total of one hundred and seven (107) patients were included in our study, including 56 men and 51 women, the male/female sex ratio was 1.1. The mean age of the patients was 57.45 years ± 15.93. The median age was 57 years [IQ = 47.5 − 69.5]. The most represented age group was 46 - 60 years, representing 37.4% of cases. Figure 1 shows the numbers according to age groups and sex.

Black race was the most common, accounting for 95.3% of patients. Other racial origins were Caucasian (1.9%), North African (1.9%), and mixed race (0.9%). Body Mass Index (BMI) was calculated for 23 patients, showing 82.6% of them with elevated BMIs.

Patients were admitted to the intensive care unit via the Emergency Department and the OBOATH’s Special Unit for the Management of SARS-CoV-2 patients, or via secondary transfers from various healthcare facilities across the country. Only four patients came directly from home.

Among the cardiovascular histories reported, high blood pressure was the main antecedent, accounting for 52.3% of patients. The research of anterior respiratory problems found asthma (2.8%) and neurologically, 6.5% of patients had a history of ischemic stroke. The antecedent of renal history showed 4.6% of patients who had chronic renal failure on dialysis. We found 1.8% of patients who were HIV-positive, one patient was pregnant, 29.0% of the patients were diabetic, and 2.8% of patients had under-gone surgery in the previous month.

Previous functional status, assessed using the Knaus scale, showed that none of our patients were bedridden prior to admission, and 60.7% of patients were receiving ongoing treatment. Patients with at least one vulnerability factor represented 73.8% of our sample. The age group in which we found the most patients with vulnerability factors was 46 - 60 years. Figure 2 shows the distribution of the most common vulnerability factors.

As prodromes, 20.60% of patients presented with flu-like symptoms and 76.60% presented with no common prodrome. Hyperthermia was sought in 69.16% of patients and found in 15.89% of patients in our sample. Asthenia was sought in 98 patients and found in 87 patients. Anorexia was sought in 96 cases and was present in 26 cases.

Cough was the main functional sign, presented by patients with 42.10% of cases, followed by myalgia (7.50%) and chest pain (3.7%). Only one patient presented with anosmia and ageusia.

Patients with respiratory distress represented 93.46% of the study population. Pulse oxygen saturation ranged from 48% to 100%, with a mean of 86.42% ± 10.47 and a median of 89 [IQ = 80 - 94]. Figure 3 represents the distribution of patients according to pulse oxygen saturation at admission.

The patients’ respiratory rate ranged from 6 to 60 cycles per minute, with a mean of 31.67 ± 9.87 and a median of 30 [IQR = 25 - 40].

Arterial blood gas analysis was performed in 26.17% of patients. Among the results obtained, 57.14% presented hypoxemia (PaO₂ < 80 mmHg) and 21.43% hypercapnia (PaCO₂ > 44 mmHg). The mean PaO₂ was 83.68 ± 37.56 mmHg and the mean PaCO₂ was 36.20 ± 11.19 mmHg.

The Liao score allowed us to assess the patients’ severity profile. The mean was 9.61 ± 3.20, for a median of 9 [IQR = 8 - 11]. The majority of patients had a critical risk profile as shown in Figure 4 .

The leukocyte count ranged from 2390 to 37,000/mm³ with a mean of 11,480/mm³ ± 6740.74 and a median of 10,000 [IQ = 6500 - 14,200]. The hemoglobin level ranged from 3.3 g/dl to 16.1g/dl with a mean of 11.40 g/dl ± 2.32 and a median of 11.8 [IQ = 10 - 13]. D-Dimers were performed in 31 patients and showed a mean of 3211.31 ng/dl ± 2911.46 and a median of 2234 [IQ = 1006.26 - 4555.01].

The paraclinical biological diagnosis of SARS-CoV-2 disease was made by performing a Rapid Diagnostic Test (RDT), serology, and/or RT-PCR. In our sample, 40.19% of patients underwent an RDT, 4.67% a serology test, and 30.84% a PCR test. Among patients who underwent a PCR test, the result was positive in 75.76% of cases.

Chest computed tomography (CT) scans were performed on 85.98% of patients in our sample, and 91.30% underwent it within 24 hours. Depending on the extent of the lesions observed, a percentage of invasion was determined by the facility’s radiologist, showing CT-scanned lung invasion greater than 50% in 78.2% of patients. CT lesions were mainly represented by ground-glass images (70.65%) and a “crazy paving” appearance in 14.13% of the chest CT scans performed. Two patients had no visible lesions ( Figure 5 ).

The SGI II score was used to assess disease severity. We found scores ranging from 6 to 82, with a mean of 31.05 ± 12.85 and a median of 29 [IQ = 22.5 - 38], representing an average predicted mortality of 12.8%. Although the majority of patients had a predicted mortality of less than 20%, the highest scores were found in the age group over 60 years.

According to the national protocol in force in Gabon, treatment began immediately, and 35.5% of patients received dual antibiotic therapy combining a third-generation cephalosporin and azithromycin. Anticoagulation was initiated in 95.3% of patients, and 85% of them were receiving curative doses of anticoagulants. None of the patients received a prescription for hydroxychloroquine.

Vascular amines were required in 3.74% of patients, and 5.61% required dialysis during their hospitalization.

The minimum length of stay for patients in the intensive care unit was 24 hours and the maximum length of stay was 32 days, with a mean of 5.93 days ± 4.95 and a median of 5 [IQR = 3 - 8]. In the study, we recorded 43.9% of deaths.

This descriptive, retrospective, single-center study aimed to describe the epidemiological characteristics of patients with coronavirus disease admitted to the intensive care unit of the Omar BONGO ONDIMBA Military Training Hospital.

Nevertheless, the results of our study must be interpreted with certain methodological limitations in mind. First, as a retrospective study, we faced constraints related to data collection. Second, our study was single-center, making it difficult to extrapolate the data collected to the general population. Finally, some variables were missing due to poor record keeping in this pandemic context with intense clinical activity during the crisis. Finally, there was sporadic inadequacy of the available technical equipment, particularly for arterial blood gas analysis and even computed tomography.

However, our study represents a significant cohort, and the data obtained deserve to be presented and compared with the literature.

The mean age was 57.45 ± 15.93 years. The most representative age group was 46 to 60 years, comprising 37.4% of patients. The results of a study conducted at the University Hospital of Libreville found a mean age of 53.5 ± 15.5 years with extremes of 16 and 90 years [9] . A thesis carried out at the Akanda Army Training Hospital highlighted a mean age of 57 ± 17 years [10] . Another Gabonese study found the same average age for patients included in general and even those admitted to intensive care units [6] . Our results are almost similar to those obtained by Donamou J. et al. at the DONKA hospital in Guinea Conakry where the mean age was 59 ± 14 years.

The results of a study conducted by the COVID-ICU group showed that the median age of patients admitted to intensive care was 63 years [11] . This median age is higher than that found in our study (57 years), and is close to that found by Gundogan et al. (67 years) [12] .

This age difference between continents is explained by a relatively young African population, unlike the European and American populations.

The study population was predominantly male with a sex ratio (M/F) of 1.1. This male predominance has been reported by several studies such as that of Gundogan et al. in Turkey (59.6% with n = 421) [12] , that of Giacomo Grasselli et al. in Lombardy (82% with n = 1591) [13] .

Studies have attempted to explain the gender difference in COVID-19 mortality by a higher expression of the ACE2 receptor that has been found in Asian men [14] .

Paradoxically, if tissue expression of ACE2 allows the virus to penetrate the cell, the soluble form of ACE2 could be a protective factor against COVID-19. Circulating ACE2 activity is indeed low in overweight or hypertensive patients, whereas it is higher in children and is positively correlated with estrogen expression [15] .

This explains, for some authors, the relative protection of children compared to adults and of women compared to men in COVID-19 [16] . This is confirmed in the study, where, even if patients under 15 years of age were not included, no patient in this age group presented to the structure with moderate or severe symptoms of SARS-CoV-2 infection.

The assessment of previous functional status using the Knaus score found 60.7% of patients with a B score, corresponding to the presence of chronic treatment. These results show that more than half of the patients had defects, which likely made them more prone to developing a severe form of COVID-19 disease.

The presence of comorbidities, such as diabetes, hypertension, heart disease, chronic lung disease and cancer, have been described as risk factors for severe disease that may lead to hospitalization in intensive care. In Gabon first, Iroungou et al. found that history of diabetes was significantly more common in patients with severe symptoms than in patients with mild symptoms and those with no symptoms (0.9%), whether these factors were considered overall or separately [6] . In Guinea Conakry, Donamou et al. found 77% of patients with at least one comorbidity [17] . This result is higher than that found by Chen et al. (50.5%) [18] Merabet et al. (69%) and Vanhems et al. (69.4%) [19] [20] but lower than that found by Gundogan et al. (87.6%) [12] , and Bay (80%) [21] . The most common comorbidities found in the series were hypertension and diabetes with 52.3% and 29.0% of patients affected respectively. Our results corroborate the data in the literature.

Multiple explanations can be put forward for the apparent association of pre-existing diabetes and the severity of COVID-19. Hyperglycemia and insulin resistance promote increased synthesis of glycosylation end products and pro-inflammatory cytokines, and this inflammatory process may be the underlying mechanism leading to a greater propensity for infection. Finally, the high expression of ACE2 in these patients may expose them to severe forms of infection [22] .

According to a systematic review with meta-analysis on the clinical characteristics of COVID-19 in China, common symptoms among ICU patients were fever (50% - 98%), fatigue (38%), cough (66% - 88%), dyspnea (63.5% - 88%), and sputum production (42%) [23] . In the study by Cummings et al. [24] in the United States, they found a predominance of dyspnea (74%), followed by cough (66%) and myalgia (26%).

In the series, we found a predominance of respiratory distress (93.46%), fatigue (81.3%), and cough (42.10%). Our results are roughly similar to the results reported by these studies and confirm the consistency of dyspnea and cough found in the literature.

COVID-19 is a progressive disease. Some patients may progress from a mild illness to a phase of worsening symptoms that can lead to ARDS or even organ failure. The occurrence of ARDS in the context of COVID-19 is widely reported. It is responsible for a mortality rate of 40 to 50% depending on the country [25] [26] . According to the medical journal The Lancet, it is the main cause of death from COVID-19 [27] . In the study, the occurrence of ARDS was confirmed in 78.5% of patients who underwent arterial blood gas analysis. This result corroborates those found in the literature: 87.3% by Grasselli et al. in Italy [25] , 76.1% by Tan et al. in a systematic review with meta-analysis [28] , 63.2% by Mitra et al. in Canada [29] .

COVID-19 is associated with several biological abnormalities. Its pathophysiology appears complex and still poorly understood. Indeed, several meta-analyses have revealed heterogeneity in the expression of the biological profile in severe COVID-19 patients [30] [31] .

In the study, we found a mean leukocyte count of 11,480 ± 6740.74/mm³. Doumbia in Mali found a mean of 13,850 ± 9360/mm³ [32] and Donamou in Conakry found hyperleukocytosis in 84% of cases [17] . This may be explained by possible bacterial superinfections, which are conducive to developing on a weakened terrain of viral pneumonia.

In the literature, D-dimer levels have been reported to be elevated in COVID-19 patients in general. A systematic review with meta-analysis published in 2020 reported that this elevation was associated with severe disease [24] . The mean D-dimer level in the study was 3211.31 ± 2911.46 ng/ml, similar to Doumbia in Mali, Gundogan et al. in Turkey, Ortiz-Brizuela et al. in Mexico, and Mitra et al. in Canada [17] [29] [32] [33] . The mechanism leading to this D-dimer elevation is not yet clearly understood. There is a dynamic correlation between elevated D-dimer levels and the occurrence of thromboembolic events in these patients [31] . In Gabon, a study focusing only on thromboembolic manifestations in patients with severe forms of SARS-CoV-2 found them in almost 11% of cases, confirming the fact that the incidence of thrombotic manifestations in patients with COVID-19 remains high, thus justifying the systematic prescription of prophylactic, or even curative, anticoagulation [8] .

RT-PCR plays a critical role in determining hospitalization and isolation of individual patients. However, its lack of sensitivity, insufficient stability, and relatively long processing time have hampered the control of the disease outbreak. In addition, a number of external factors may affect the results of RT-PCR tests, including sampling operations, sample source (upper or lower respiratory tract), sampling time (different period of disease development), and detection kit performance. As such, RT-PCR test results should be interpreted with caution [34] .

In our sample, less than half of the patients had received a PCR test. A positive PCR result was found in 75.76% of cases, when it was performed. In another study, a cohort in America, patients were repeatedly negative on RT-PCR for SARS-CoV-2, despite presenting a range of typical signs and/or symptoms associated with COVID [24] . Furthermore, another retrospective Chinese study that aimed to compare the sensitivity and specificity of RT-PCR and chest CT, revealed that the sensitivity of chest CT is estimated at 97%, and the specificity at 25%, compared to a sensitivity of 65% and a specificity of 83% for RT-PCR [35] . These results suggest that a negative RT-PCR test should not preclude COVID treatment in the presence of clinical suspicion of the disease, which was done in the study, clinical and CT parameters were considered sufficient to retain the diagnosis.

Chest CT is a conventional, non-invasive imaging modality with high accuracy and speed. It is considered a screening and early diagnostic tool in COVID-19 pneumonia [35] .

In this series, CT was performed in 85.98% of patients. Ground-glass appearance was the most common radiological lesion at 70.65%, followed by crazy paving, which represented 14.13%. These results are similar to those of Marc Garnier et al., in France, who found ground-glass appearance in 68 to 83% of cases, followed by crazy paving lesions in 15% to 34% [36] . Extension of lesions on chest CT to more than 75% was found in 14.13% of patients. The main CT sign of severity is the extent of parenchymal abnormalities on the initial CT scan. Many studies report a correlation between the extent of lesions and clinical severity [37] - [39] .

To date, no specific treatment has been recommended for coronavirus infection. Treatment varies from one country to another depending on the established national protocol. Data from the literature and the WHO agree that corticosteroids, particularly dexamethasone, reduce the number of deaths in intensive care [40] . Other treatments such as antivirals, antibiotics and anticoagulants vary in use from one country to another. In Gabon, the management of severe forms of covid-19 has been codified and a protocol has been implemented by the Coronavirus Monitoring and Response Committee (COPIL) which systematically combines: oxygen therapy, antibiotic therapy, corticosteroid therapy, anticoagulation and vitamin therapy [41] .

The management of patients with respiratory failure due to SARS-CoV-2 pneumonia focuses on respiratory support, using several modalities that constitute the corner-stone of treatment for acute respiratory distress syndrome [42] .

In the study, 45.79% of patients received conventional oxygen therapy, 43.93% received non-invasive ventilation (NIV), and 6.42% received mechanical ventilation.

In our series, antibiotic therapy was routinely used in accordance with the current national protocol [41] . This consisted of dual antibiotic therapy combining ceftriaxone and azithromycin. A fluoroquinolone could be added, along with other molecules, in case of obvious signs of bacterial superinfection or in patients receiving mechanical ventilation.

This protocol is different from that of Morocco which recommends the combination of Hydroxychloroquine-Azithromycin as a first-line treatment, then antivirals as a second line. Antibiotics are only recommended if there is evidence of bacterial superinfection [43] .

The risk of thromboembolic events in COVID-19 patients admitted to intensive care appears very high due to the presence of risk factors, including age, comorbidities, prolonged bed rest, and the pathophysiology of the disease. In an update published by Satre Buisson in September 2020, anticoagulant treatment for prophylactic purposes or at higher or even curative doses, taking into account the risk of bleeding, is widely recommended in patients with COVID-19, in order to limit the impact of hypercoagulability induced by SARS-CoV-2 infection [44] . In this series, the majority of patients received curative doses of anticoagulation in the absence of contraindications.

Circulatory failure reflects the critical condition of our patients, resulting in the transition to septic shock. The management of septic shock in patients with COVID-19 is not specific and strictly follows the recommendations of the Surviving Sepsis campaign [45] . In the series, we used vasoactive amines in 3.74% of patients. In France, during the first 3 years of the epidemic, 27% of patients required the administration of catecholamines [46] .

The average length of hospitalization of patients was 5.93 days. This result is close to that found in France where the average length of hospitalization was 7 days [46] . Lower than that reported in China (10 days) [47] and in Conakry (12 days) [17] . The length of hospitalization varies depending on the severity of the patients’ condition but also on the care methods specific to each healthcare establishment which may or may not extend the stay of its patients.

Literature data have shown high mortality among patients with severe forms of COVID-19. In terms of numbers, this mortality appears quite heterogeneous. A mortality rate of 15.4% has been reported in Vancouver, 25% in Conakry, 55.1% in Turkey, and 85% in Algeria [12] [17] [19] [29] . In the study, the mortality rate was 43.93%. This heterogeneity may reflect the heterogeneity of patients based on their age, comorbidity, superinfection rate, and treatment time.

The SARS-CoV-2 disease, when it emerged in November 2019, plunged the world into a veritable health crisis. Although its clinical presentation is polymorphic and despite numerous limitations, including significant missing data, this study allowed us to better understand the typology of patients admitted to intensive care of Libreville for the management of SARS-CoV-2 pneumonia.

The majority of patients were men, between 46 and 60 years of age with comorbid hypertension. Admission to intensive care occurred due to respiratory distress, with imaging findings of lesions typical of SARS-CoV-2 pneumonia on 75% of the CT scans.

To date, there is no evidence of effective treatment for COVID-19. Treatment is symptomatic, focusing on respiratory support, corticosteroid therapy, antibiotic therapy, and heparin therapy. Since the start of vaccination, cases of COVID-19 appear to be declining. This led to the WHO lifting the global health emergency on May 5, 2023, while reiterating that the virus has neither been eradicated nor rendered harmless. We must, therefore, remain cautious by maintaining barrier measures while preparing response programs to deal with other possible epidemics due to the emergence of a new variant. The evidence is that the endemic nature of the disease is proven, with the admission of an average of 5 to 6 patients with severe COVID-19 per year to our facility in 2024.

Finally, with the aim of improving and anticipating patient care, similar studies in other university hospitals in the capital should be conducted to support the epidemiological findings and determine the risk factors associated with severe forms of COVID-19 specific to the Gabonese population.