Epidemiological, Clinical, Etiological and Evolutionary Profile of Patients Hospitalized for Upper Digestive Hemorrhage at the Sourô Sanou University Hospital Center in Bobo-Dioulasso ()
1. Introduction
Upper gastrointestinal bleeding (UGIB) is a common and potentially life-threatening medical and surgical emergency requiring rapid and effective management. It represents a major cause of morbidity and mortality worldwide, with epidemiological and clinical profiles that vary considerably depending on the region [1]-[3].
In developed countries, particularly in Europe and the United States, the incidence of UGIB varies from 45/100,000 to 150/100,000 inhabitants, depending on the country, with a decline in incidence reaching 40% in some countries [4]. However, despite progress in the practice of therapeutic endoscopy, mortality remains high, reaching 5% to 10% [5] and even 14% in 17.4% in Ghana [6].
In Africa, UGIB is often associated with higher mortality rates due to its occurrence in conditions of inaccessibility to therapeutic endoscopy and underlying pathologies [7]. Varying hospital prevalences of 7.3% in Togo [8], 6.7% in Mali [9], and 24.5% in Uganda [10], with respective mortality rates of 5%, 14.3% and 1.6% have been reported.
In Burkina Faso, Sombié [2] in 2015 reported that UGIB of ulcerative origin represented the most frequent etiology with 52.3% of cases and a mortality of 17% at the Yalgado Ouédraogo University Hospital Center. A hospital frequency of 11.5% and an overall mortality of 12.8% were reported by Kinda in 2021 at the Sourô Sanou University Hospital in Bobo-Dioulasso in a study on digestive hemorrhages [11]. Series from developed European countries are characterized by cohorts of elderly patients and, with multiple comorbidities weakening the terrain and favoring death [12]. While studies by African authors reported younger subjects rather weakened by bacterial infections, an unstable hemodynamic state, and hepatocellular insufficiency, factors favoring their death [2] [13]-[16]. An understanding of the conditions under which this high mortality reported in studies occurs is necessary to improve care. The objective of this study was to describe the epidemiological, clinical, etiological, and evolutionary profile of patients hospitalized for upper gastrointestinal bleeding at the Sourô Sanou University Hospital in Bobo-Dioulasso, in order to better guide therapeutic interventions and prevention.
2. Patients and Methods
2.1. Study Framework
The study was conducted in the Hepato-Gastroenterology (HGE) Department of the Sourô Sanou University Hospital in Bobo-Dioulasso. The Sourô Sanou University Hospital is the reference center for western Burkina Faso, serving the health regions of Hauts-Bassins, Cascades, Sud-Ouest, and Boucle du Mouhoun.
2.2. Study Type and Period
This was a retrospective cohort study conducted from January 1, 2021, to December 31, 2023, a study period of 36 months.
2.3. Study Population and Sampling
The target of our study was all patients with upper gastrointestinal bleeding in the Sourô Sanou University Hospital catchment area. The source population consisted of patients hospitalized in the HGE department during the period. Patients admitted to the Sourô Sanou University Hospital HGE Department for UGIB during the study period and with complete, usable clinical records were included. Patients admitted to the Hepato-Gastroenterology department of the Sourô Sanou University Hospital Center for UGIB during the study period and with complete, usable clinical records were included.
Sampling consisted of a census and exhaustive recruitment. All patients meeting the study criteria were included.
2.4. Study Variables
The study variables were grouped into five main items: sociodemographic data (age, sex, residence, and occupation), clinical data (mode of admission, patient history, methods of UGIB detection, physical signs upon admission, and presence or absence of shock upon admission), paraclinical data (hemoglobin level, serum creatinine level, and upper gastrointestinal endoscopy results), therapeutic data (initial nonspecific treatment, specific medical treatment, surgical treatment), and outcome data (length of hospitalization, mode of discharge, existence or absence of recurrence, and occurrence or absence of complications during hospitalization).
2.5. Data Collection Techniques and Tools
Data collection was carried out using a form developed with Epi-Info and consisted of a review of the hospitalization register and clinical records of patients hospitalized in the Hepato-Gastroenterology department of the Sourô Sanou University Hospital during the study period.
2.6. Statistical Data Analysis
Data entry and statistical analysis were performed using Microsoft Office 2016 and STATA version 14. Quantitative variables were measured by their means or medians, along with their standard deviations, and qualitative variables by their proportions.
2.7. Ethical Considerations
The study received approval from representatives of the Sourô Sanou University Hospital Institutional Ethics Committee before its initiation.
Data collection was conducted with respect for patient anonymity and confidentiality. Therefore, only the medical record number was recorded. No patient names or identifying information were included. These data were used solely for this study.
3. Results
During the study period, 1791 patients were hospitalized in the Sourô Sanou University Hospital HGE department, including 303 patients for UGIB, representing an overall frequency of 16.9%. The usable records of 291 patients were included, with variations in frequency from one year to the next. Patient flow diagram and frequency of UGIB are detailed in Figure 1.
Figure 1. Patient flow diagram and frequency of upper gastrointestinal bleeding.
3.1. Sociodemographic Characteristics
In our series, the mean age was 50 years ± 1.02 years, with a range of 36 to 63 years, and the sex ratio was 2.6. The various sociodemographic characteristics of patients are detailed in Table 1.
Table 1. Sociodemographic characteristics of patients (N = 291).
Variables |
Number |
Percentage |
Gender |
|
|
Male |
210 |
72.2 |
Female |
81 |
27.8 |
Main Occupation |
|
|
Informal Sector |
150 |
51.6 |
Farmer |
86 |
29.6 |
Employee |
31 |
10.6 |
Housewife |
17 |
5.8 |
Pupil/Student |
7 |
2.4 |
Place of Residence |
|
|
Urban |
161 |
55.3 |
Rural |
130 |
44.7 |
3.2. Patient Medical History and Lifestyle
A history of peptic ulcer disease was found in 58 (20%) patients. Approximately 39 (13%) cases of viral hepatitis were documented, including 27 hepatitis B virus (HBV) infections, 10 hepatitis C virus (HCV) infections, and 2 HBV and HCV coinfections. Hypertension and cirrhosis affected 12% and 10% of patients, respectively.
Regarding risk behaviors, alcohol consumption, smoking, and nonsteroidal anti-inflammatory drug (NSAID) use were reported in 65 (23%), 45 (16%), and 35 (12%), respectively, of patients with gastroduodenal peptic ulcer.
3.3. Clinical and Laboratory Data
The majority of patients, 255 (87.6%), were evacuated in an emergency setting, and the others were seen during an outpatient consultation before hospitalization. The main mode of presentation was hematemesis (82%), followed by melena (54%); other functional signs such as abdominal pain and dizziness were reported in 108 patients (37%). Signs of hemodynamic shock were reported in 7.2% of patients. The mean complete Rockall score, assessed in 162 patients, was 2.8 ± 1.4, with a range of 0 to 8 (Table 2).
Table 2. Distribution of the various clinical and laboratory signs of the patients.
Variables |
Number |
Percentage |
Normal Consciousness |
|
|
Yes |
269 |
92.4 |
No |
22 |
7.6 |
Paleness |
|
|
Yes |
191 |
65.6 |
No |
100 |
34.4 |
Jaundice |
|
|
Yes |
36 |
12.4 |
No |
255 |
87.6 |
Ascites |
|
|
Yes |
33 |
11.3 |
No |
258 |
88.7 |
Hepatomegaly |
|
|
Yes |
40 |
13.8 |
No |
251 |
86.2 |
Oedema of the Lower Limbs |
|
|
Yes |
36 |
12.4 |
No |
255 |
87.6 |
Normal Pulse Rate |
|
|
Yes |
176 |
60.5 |
No |
115 |
39.5 |
Blood Pressure |
|
|
Normal |
43 |
14.9 |
Low |
198 |
68.1 |
High |
49 |
17.0 |
Hypovolemic Shock |
|
|
Yes |
21 |
7.2 |
No |
270 |
92.8 |
Hemoglobin Level (g/dl) |
|
|
<7 |
155 |
53.4 |
[7 - 10] |
81 |
27.8 |
]10 - 14] |
37 |
12.8 |
>14 |
18 |
6.0 |
Hematocrit Level (%) |
|
|
<20 |
134 |
46.2 |
[20 - 30] |
98 |
33.5 |
]30 - 45] |
50 |
17.3 |
>45 |
9 |
3.0 |
Rockall Score |
|
|
[0 - 2] |
82 |
50.6 |
<2 |
80 |
49.4 |
3.4. Etiologies of Upper Gastrointestinal Bleeding
In our study, 162 (55.7%) patients underwent upper gastrointestinal endoscopy. The main causes of upper gastrointestinal bleeding were gastroduodenal peptic ulcer (43.2%), followed by ruptured esophageal varices (19.1%). Table 3 below provides a breakdown of UGIB etiologies based on upper gastrointestinal endoscopy results.
Table 3. Distribution of the main etiologies of upper gastrointestinal bleeding among patients who underwent upper gastrointestinal endoscopy.
Variables |
Number |
Percentage |
Gastroduodenal peptic ulcer |
70 |
43.2 |
Rupture of esophageal varices |
31 |
19.1 |
Gastropathies |
23 |
14.2 |
Gastric dysmitosis |
3 |
1.9 |
Mallory-Weiss syndrome |
2 |
1.2 |
Normal upper gastrointestinal endoscopy |
33 |
20.4 |
Total |
162 |
100 |
*A patient could have several types of lesions.
3.5. Therapeutic Data
Initial management of unstable patients was provided in the Sourô Sanou University Hospital emergency department, and patients were transferred to the HGE after stabilization. However, additional resuscitation measures, such as fluid replacement and blood transfusion, are sometimes performed in the HGE department following possible secondary decompensation (Table 4).
Table 4. Distribution of the main treatments administered to hospitalized patients (N = 291).
Variables |
Number |
Percentage |
Vascular Filling |
|
|
No |
269 |
92.4 |
Yes |
22 |
7.6 |
Blood Transfusion |
|
|
No |
156 |
53.6 |
Yes |
135 |
46.4 |
Oxygen Therapy |
|
|
Yes |
13 |
4.5 |
No |
278 |
95.5 |
Proton Pump Inhibitors |
|
|
Yes |
277 |
95.2 |
No |
14 |
4.8 |
3.6. Evolutionary Data
The median length of hospitalization was five days, with an interquartile range of 3 to 8 days. The majority of patients were discharged alive, and 58 died, representing a case fatality rate of 20%. Table 5 shows the distribution of patients by discharge method.
Furthermore, 10% (n = 29) of cases of recurrence were reported, and the major complication during hospitalization was decompensated anemia in 75.9% of cases (n = 22). Other complications were renal failure and sepsis in 4 and 3 cases, respectively.
Table 5. Distribution of discharge methods for different patients.
Variables |
Number |
Percentage |
Alive |
166 |
57 |
Death |
58 |
20 |
Discharge against medical advice |
46 |
16 |
Transfer |
15 |
5 |
Discharge without medical advice |
6 |
2 |
Total |
291 |
100 |
In the bivariate analysis, sociodemographic factors and medical histories associated with mortality (p < 0.2) included age over 25 years, viral hepatitis, cirrhosis, and nonsteroidal anti-inflammatory drug use. Clinical and paraclinical factors associated with death included the presence of jaundice, edema, ascites, and hepatomegaly (p < 0.2). Therapeutic and progressive factors associated with mortality were an underlying bacterial infection, shock, and length of hospitalisation.
4. Discussion
Our study presented some limitations related to the lack of archiving and the handwritten system still in use. This may have led to a lack of completeness in hospitalized cases and the variables recorded in medical records and consultation registers. However, this study also contributes to knowledge on digestive hemorrhages in resource-limited countries in sub-Saharan Africa, and the results obtained are worth discussing.
4.1. Hospital Frequency of UGIB
The frequency of gastrointestinal bleeding (16.9%) during our study was comparable to that found by Kinda [11] in 2021 in Bobo-Dioulasso (17.9%). However, it remains higher than that reported in the Yahya study in 2022 in Kaduna [17], but lower than those reported by Saba [18] in 2019 in Ouagadougou (27.6%) and by Opio [10] in 2022 in Amolatar (24.5%). This difference in frequency could be explained by improved access to care through decentralized management of certain cases in district hospitals and private health centers. This thus limits medical evacuations to tertiary centers or referral hospitals.
The prevalence is clearly decreasing in developed countries due to better control of viral hepatitis through universal vaccination against hepatitis B and effective pangenotypic treatment of hepatitis C, thus reducing the progression to cirrhosis and its complications [19]. In sub-Saharan Africa [20], in several countries, we also noted a slight decrease in the frequency of UGIB reported during the study period, probably reflecting the overall improvement of health and care systems, with more specialists, allowing better diagnosis. Despite this decrease, the frequency of UGIB still remains high, probably related to self-medication by populations with NSAIDs, and the still high prevalence of viral hepatitis and cases of cirrhosis, sources of Esophageal varices [8].
4.2. Epidemiological Characteristics
In our study, the median age of patients was 50 years. Similar results have been reported by some authors in Africa, with median ages ranging from 42 to 54 years [14] [19] [21] [22]. This could rather be explained by the causes of UGIB in Africa, which are gastroduodenal peptic ulcer and rupture of esophageal varices [17]. Indeed, gastroduodenal peptic ulcer is a condition of young adults (mean age 46.2 years) that matures around the age of fifty. The rupture of esophageal varices is a corollary of the high prevalence of chronic viral hepatitis, of which cirrhosis is one of the main complications, also occurring at a young age. The average ages reported by authors in the West [23] are over 60 years. This reflects the aging of their populations, often associated with comorbidities and the regular use of NSAIDs and aspirin. Our study reports a male predominance with a sex ratio of 2.6, consistent with literature data [2] [19] [24]. There is a link between male predominance and the etiologies of UGIB; indeed, in the literature, gastroduodenal peptic ulcer is a condition that mainly affects the male gender, more exposed to certain risk factors in our context (alcoholism, smoking); the same is true for cirrhosis [2] [10] [20].
4.3. Clinical and Paraclinical Characteristics
A history of cirrhosis was reported in 10% of our patients. Surial et al. [25], through a 2021 systematic review, found a prevalence of viral cirrhosis between 4% and 13%. Indeed, chronic carriage of hepatitis B and C viruses is the main cause of cirrhosis in Burkina Faso [15]. NSAID use was found in 12% of patients. Proportions of 16% and 22% were reported by Gassaye [26] and Kinda [11], respectively. The retrospective nature of our study, the urgent nature of gastrointestinal bleeding, and self-medication by often uneducated populations, unaware of the types of NSAID molecules used, could lead to underreporting of NSAID use history in our patients.
Hematemesis was the main mode of revelation in our series, with 82% of cases. Hematemesis remains the most frequently reported mode of exteriorization of UGIB in most African studies [2] [8] [11] [19]. Signs of hemodynamic shock were reported in 7.2% of patients. Obeidat et al. [27], in a meta-analysis, reported a proportion of hemodynamic instability between 22% and 25% at the admission of patients and less than 7% in hospitalized patients, which is comparable to our result. These results are explained by the fact that patients are hemodynamically stabilized upstream, in the emergency room or in intensive care, before their hospitalization in specialized services. The hemoglobin thresholds that triggered blood transfusion were those of less than 7 g/dL and <10 g/dL for patients with heart disease or renal failure. Vascular replacement was performed in the presence of signs of shock and unstable hemodynamic status (tachypnea, pallor of the mucous membranes, cold extremities, thready pulse, tachycardia, arterial hypotension), with or without an altered state of consciousness, while awaiting blood transfusion. Bhuyan et al. [28] reported a frequency of 50% in a population of elderly people with more comorbidities. In our series, 53.4% of patients had a hemoglobin level below 7 g/dl. Our result is similar to that of Bignoumba [14] in Gabon (50%) and higher than that of Moussa [16] in Tchad (30%). This could be explained, on the one hand, by the delay in consultation due to the low socioeconomic level and, on the other hand, by the difficulty of transfusion at the peripheral level linked to the insufficiency of labile blood products.
In our study, 162 (55.7%) patients underwent upper gastrointestinal endoscopy. During the study, 45% of eligible patients did not undergo endoscopy for several reasons: geographic inaccessibility (unavailability of digestive endoscopy at this hospital at the time the study was conducted), financial inaccessibility due to lack of social security, and third-party payment. And some patients also died before stabilization of their hemodynamic status. This missing information may indeed have a potential impact on the reported etiological distribution. However, the arguments concerning the youth of the population and the tendency toward self-medication with NSAIDs are also valid regarding the highlighted distribution. The two most common etiologies were gastroduodenal ulcer (43.2%) and ruptured esophageal varices (19.1%). Our result overlaps with those found by Kinda [11] and Sombié [2] in Burkina Faso, Bagny [8] in Togo, and Chaabane [29] in Tunisia, who all reported in their studies that gastroduodenal ulcerative disease is the main etiology of UGIB. This high frequency could be justified by a high prevalence of Helicobacter pylori infection in Africa, often asymptomatic, affecting more than 50% of the population [30], and the over-the-counter sale of NSAIDs, encouraging self-medication. As for the rupture of esophageal varices, it is also reported in second place in several studies [14] [31] from countries where the prevalence of cirrhosis of viral origin is also high. Other studies have reported the predominance of ruptured esophageal or gastric varices [3] [17] over gastric or duodenal ulcers. In all cases, these two etiologies are the most frequently implicated.
4.4. Therapeutic and Progression Characteristics
Nearly half of our patients (46.4%) received a blood transfusion. Higher proportions of transfusion cases were reported by Kinda [11], Ntagirabiri [31], and Dicko [9], in 63.4%, 63.9%, and 61.9% of cases, respectively. Improvements in the labile blood product supply system, with the involvement of the highest government authorities, allowing rapid transfusion of patients in the emergency department, could explain the rate reported in our study. During hospitalization, 95.2% of patients (n = 277) received proton pump inhibitor therapy. These results thus corroborate the predominance of gastroduodenal ulcer, identified as the main etiology of UGIB in our series. No cases were treated with vasoactive substances or required hemostatic surgery. It is important to note the geographical and financial inaccessibility of vasoactive drugs and the unavailability of therapeutic and interventional endoscopy equipment at the Sourô Sanou University Hospital.
Bleeding recurrence during hospitalization was observed in 10% of patients. Lower results were reported by El Mekkaoui [32] in Tunisia (7.5%) and Dicko [9] in Mali (7.9%). The frequency of recurrence of UGIB depends on the underlying pathology; it is frequent in cases of rupture of esophageal varices and depends on the quality of care [15]. The hospitalization time observed in this study is superimposable with those of the studies of Kinda [11] and of Ntagirabiri [31], which reported an average stay of 6 days. Early performance of a digestive endoscopy in cases of upper gastrointestinal bleeding, as soon as the patient’s hemodynamic state has stabilized, allows for rapid and appropriate intervention and is thus associated with a reduction in hospitalization times and rebleeding.
During the hospital stay, 58 cases of death, representing a mortality of 20%, were observed. A high mortality rate of between 14% and 22% has been reported by other authors in Sub-Saharan Africa [8] [9] [11] [15] [13] [31] contrasting with the results of series from developed countries [19] [29] [32] [33] which reported lower mortality rates of between 1.9% and 5.5%. The delay in consultation and treatment, the absence of a technical platform for endoscopic hemostasis of ulcers, and the failure to perform ligation of esophageal varices in our study could explain this high mortality. The mean complete Rockall score was 2.8, a result close to those reported by Sombié [2] and Zombré [34], which were 3.2 and 4.7, respectively. The Rockall score allows for the assessment of the risk of death. However, the incompleteness of the data due to a lack of endoscopic results in some patients limits its interpretation in our context.
In the bivariate analysis, sociodemographic factors and medical histories associated with mortality (p < 0.2) included age over 25 years, viral hepatitis, cirrhosis, and nonsteroidal anti-inflammatory drug use. Clinical and paraclinical factors associated with death included the presence of jaundice, edema, ascites, and hepatomegaly (p < 0.2). Therapeutic and progressive factors associated with mortality were an underlying bacterial infection, shock, and length of hospitalisation. In the Opio study, the factors significantly associated with UGIB on bivariate analysis included sex, previous gastrointestinal tract bleeding, smoking, alcohol consumption, engagement in fishing, schistosomiasis, esophagitis, and tuberculosis disease. Alcohol use (p = 0.030) and having a previous gastrointestinal bleeding (p < 0.001) were the factors independently associated with UGIB in the study of Opio [10]. Most acute upper gastrointestinal bleeding occurs in people who are already predisposed, which could explain this situation. Indeed, people who have already experienced gastrointestinal bleeding already have pre-existing risk factors, generally chronic, making them vulnerable and exposed to an increased risk of recurrence.
5. Conclusion
Upper gastrointestinal bleeding remains common in our context. It mainly affects young adult males. It most often manifests as hematemesis, and the main cause is peptic ulcer disease. Management does not benefit from any endoscopic treatment or vasoactive drugs, which is a poor prognostic factor. Mortality remains high, and the factors associated with this mortality were the variceal origin of the bleeding and underlying bacterial infections. Effective control of viral hepatitis B and C, which cause cirrhosis, and accessibility to vasoactive drugs and endoscopic hemostasis treatments could improve patient prognosis.