Probabilistic Antibiotic Therapy in the Infectious Diseases Department of the Yalgado Ouédraogo University Hospital (CHU-YO) in Ouagadougou, Burkina Faso ()
1. Introduction
Antibiotic resistance remains a major public health problem, particularly in developing countries such as Burkina Faso, where hygiene conditions are still precarious and where the use of antibiotics is often abused and poorly controlled [1]. Rational use of antibiotics is a crucial parameter of quality of care. However, in Burkina Faso, as in most resource-limited countries, limited access to confirmatory diagnostic tests forces practitioners to resort frequently to probabilistic antibiotic therapy. This prescription is based on clinical examination and knowledge of the bacterial ecology of the health facility in question. It does not take into account either the etiological diagnosis or the sensitivity of the bacteria responsible for the infection. Although this practice has demonstrated its impact on improving the vital prognosis of patients, it exposes them to selection pressure for resistant bacteria [2] [3] [4] [5]. Studies conducted in Burkina Faso have shown high rates of probabilistic antibiotic prescription. Youl et al. had already observed that ceftriaxone was prescribed probabilistically in almost all patients (98.2%), and that this prescription was inappropriate in 29.5% of cases [6]. The aim of our study, which takes place some ten years later, is to determine the extent of probabilistic antibiotic therapy.
2. Patients & Methods
The study took place in the infectious diseases department of the CHU Yalgado Ouédraogo. This department is specialized in the management of infectious diseases with or without epidemic potential. This was a cross-sectional study with retrospective data collection of patients hospitalized in the infectious diseases department during the period from January 1, 2005 to December 31, 2020. Our study included patients who were hospitalized in the infectious diseases department of the YO University Hospital for a bacterial infectious disease and who had received a probabilistic anti-antibiotic treatment. Patients whose records were not usable were excluded from our study.
Data were collected using a collection form; sociodemographic, clinical, therapeutic, and evolutionary variables were collected from hospitalization records.
The sampling was exhaustive and the variables sought were sociodemographic data (age, sex, residence, profession), clinical data (functional signs, general signs, physical signs), the presumptive diagnosis that led to probabilistic antibiotic therapy, the antibiotics used, the duration of hospitalization, and the evolution under probabilistic treatment.
The data collected were entered, processed and analyzed on a microcomputer using SPSS, Word and Excel software version 2016.
Ethically, the confidentiality and anonymity of the patients were respected in the processing and analysis of the data. We obtained authorization for data collection from the Director General of the CHU-YO.
Operational definitions:
- Fairly good general condition: patient with a general condition stage I or stage II according to the WHO classification.
- Poor general condition: patient with a general condition stage III or stage IV according to the WHO classification.
- Presumptive diagnosis: diagnosis suspected on the basis of clinical examination
- Probabilistic antibiotic therapy: any antibiotic therapy administered before confirmation of the diagnosis.
- Duration of treatment: the number of days between the beginning and the end of the probabilistic treatment administered.
3. Results
3.1. Sociodemographic Aspects of the Patients
A total of 330 patients had received probabilistic antibiotic therapy. The male sex was the most represented with 53% or a sex ratio of 1.12. The following Table 1 shows the distribution of patients in our study according to age groups.
![]()
Table 1. Distribution of patients by age group.
The age range of 20 to 40 years was the most represented (42%). The average age of the patients was 33 ± 14 years. The majority of patients worked in the informal sector (49%). More than half of the patients lived in urban areas (56.36%).
3.2. Clinical Aspects
The history of our patients is presented by HIV infection found in 15% of patients, hypertension in 5%, sickle cell disease in 2%, diabetes (1%) and hepatitis B (1%). Clinically, the majority of patients (57%) had a poor general condition on admission to the infectious diseases department. The reasons for consultation were dominated by fever, alteration of general condition, neurological disorders (headache, consciousness disorders, convulsions), digestive disorders (vomiting, diarrhea, abdominal pain), respiratory signs (dyspnea, chest pain, cough), urinary signs (dysuria, pollakiuria and micturitional burning) and diffuse pain. The physical examination showed that 48.1% of the patients had meningeal irritation syndrome, 10% had convulsions and 10% had focal signs, trismus was present in 4% of the patients and facial paralysis in 3%. In the digestive system, hepatomegaly was present in 29% of patients and digestive candidiasis in 31%. Respiratory examination showed crepitus and fluid effusion syndrome in 26.83% and 20.62% of patients respectively.
3.3. Presumptive Diagnosis
The presumptive diagnosis was dominated by meningitis, salmonellosis, and pneumonia as illustrated in Table 2.
![]()
Table 2. Distribution of patients by presumptive diagnosis.
3.4. Therapeutic Aspects
In our study, 67% of patients had received probabilistic antibiotic therapy. Betalactam was the most used antibiotic family (51.28%) in probabilistic antibiotic therapy. The most prescribed molecule was ceftriaxone (38.61%). Aminosides were the second most prescribed antibiotic family (14.05%) with gentamicin being the most prescribed (13.26%). Quinolones were the 3rd most prescribed (11.48%) with mainly ciprofloxacin being prescribed. Table 3 shows the distribution of antibiotics used by class.
![]()
Table 3. Frequency of prescription of molecules by antibiotic family in probabilistic treatment.
Monotherapy was prescribed in 59% of patients. Dual therapy was prescribed in 23% of patients; the combination of ceftriaxone + gentamycin was the most commonly used (41%) for dual antibiotic therapy; triple therapy was prescribed in 10% of patients and the combination of ceftriaxone + metronidazole + gentamicine was the most commonly prescribed; quadruple therapy (Isoniazid, Ethambutol, Rifampicin, and Pyrazinamide) was prescribed in 8% of patients with tuberculosis.
The injectable route was mostly used for probabilistic antibiotic therapy (86% versus 15% for the oral route). The duration of probabilistic antibiotic therapy varied from 5 days to 30 days. The evolution under treatment was marked by the death of 50 patients, i.e. a lethality of 15% and 1.6% of patients were discharged against medical advice.
4. Discussion
Our study had limitations and constraints. Indeed, the sample of our study was not representative enough. Some files were not complete and other files were missing. Despite these limitations, our results give rise to some comments and discussions.
Infectious diseases are a frequent reason for hospitalization in the infectious diseases department of the Yalgado Ouédraogo University Hospital. They lead to probabilistic and sometimes inappropriate antibiotic prescriptions [6] [7] [8] with a high risk of emergence of resistant bacteria. In our study, 67% of prescriptions were probabilistic. This rate is comparable to those found by Kouanda and Fongang in Burkina and Mali respectively [6] [9] [10]. However, in the multicenter study by Michard et al., lower rates of probabilistic antibiotic therapy were observed in Mauritania, Niger and Senegal [11]. In European series, this rate varied from 43% to 70% depending on the study [12] [13] [14]. The results of the studies mentioned above confirm, to varying degrees, the frequency of antibiotic prescription for treatment in hospitals [11] [15]. In our series, beta-lactams were the most prescribed. This can be explained by the high number of cases of meningitis diagnosed during the epidemic periods that Burkina Faso has experienced. Indeed, during meningitis epidemics, the Burkina Faso government provided ceftriaxone free of charge for the management of cases. This justifies, among other things, the important use of this molecule. Other reasons for the frequent use of this molecule are its relatively affordable cost and accessibility [6]. Dual antibiotic therapy was the dominant prescription in our series as in the series of Fongang et al. [7]. The combination of ceftriaxone + gentamycin was the most prescribed combination in our series as in most African series [6] [7] [8]. However, in the study by Youl et al. ceftriaxone was frequently combined with metronidazole [4].
5. Conclusion
Antibiotic therapy is a common practice in the infectious diseases department of the Yalgado Ouédraogo University Hospital. Even if it improves the outcome of patients, it exposes to the selection of resistant mutants. Monitoring the efficacy of ceftriaxone, gentamicin and ciprofloxacin is essential given the extent of the prescription of these molecules for probabilistic treatment. This follow-up must also concern animals according to the “One Health” approach. To do this, it is important to improve the technical platform of laboratories.