TYPE Original Research

PUBLISHED 16 January 2023

DOI 10.3389/fpubh.2022.1088590

Assessment of blood cultures and

OPEN ACCESS antibiotic susceptibility testing for
bacterial sepsis diagnosis and
EDITED BY
Kenneth Iregbu,
University of Abuja, Nigeria

REVIEWED BY
Ifeyinwa Nwafia,
utilization of results by clinicians in
University of Nigeria Teaching Hospital Ituku-
Ozalla, Nigeria
Claudio Farina,
Benin: A qualitative study
ASST Papa Giovanni XXIII, Italy
Samuel Taiwo,
Ladoke Akintola University of Brice Boris Legba1 , Victorien Dougnon1*, Hornel Koudokpon1 ,
Technology, Nigeria
Zhihui Zhou,
Sointu Mero2,3 , Riku Elovainio3,4 , Matti Parry3,5 , Honoré Bankole1 and
Sir Run Run Shaw Hospital, China Kaisa Haukka3,6
Shaheen Mehtar,
Infection Control Africa Network, South Africa 1
Research Unit in Applied Microbiology and Pharmacology of Natural Substances, Research Laboratory in
Applied Biology, Polytechnic School of Abomey-Calavi, University of Abomey-Calavi, Abomey-Calavi, Benin,
*CORRESPONDENCE 2
Human Microbiome Research Program, Faculty of Medicine, University of Helsinki, Helsinki, Finland,
Victorien Dougnon 3
Physicians for Social Responsibility, Helsinki, Finland, 4 Tampere Center for Child, Adolescent, and Maternal
[email protected]
Health Research (TAMCAM): Global Health Group, University of Tampere, Tampere, Finland, 5 New Children’s
SPECIALTY SECTION Hospital, University of Helsinki, Helsinki, Finland, 6 Department of Microbiology, University of Helsinki,
This article was submitted to Helsinki, Finland
Infectious Diseases: Epidemiology and
Prevention,
a section of the journal
Frontiers in Public Health
Objectives: We assessed the current status of blood culture and antibiotic
susceptibility testing (AST) practices in clinical laboratories in Benin, and how the
RECEIVED 03 November 2022
ACCEPTED 28 December 2022 laboratory results are used by physicians to prescribe antibiotics.
PUBLISHED 16 January 2023 Methods: The qualitative study covered twenty-five clinical laboratories with a
CITATION bacteriology unit and associated hospitals and pharmacies. Altogether 159 laboratory
Legba BB, Dougnon V, Koudokpon H, Mero S,
Elovainio R, Parry M, Bankole H and Haukka K
staff, physicians and pharmacists were interviewed about their perceptions of the state
(2023) Assessment of blood cultures and of laboratory diagnostics related to sepsis and the use of antibiotics. Face-to-face
antibiotic susceptibility testing for bacterial interviews based on structured questionnaires were supported by direct observations
sepsis diagnosis and utilization of results by
clinicians in Benin: A qualitative study.
when visiting five laboratories in across the country.
Front. Public Health 10:1088590. Results: Only 6 laboratories (24%) conducted blood cultures, half of them
doi: 10.3389/fpubh.2022.1088590
with a maximum of 10 samples per month. The most common gram-negative
COPYRIGHT bacteria isolated from blood cultures were: Escherichia coli, Salmonella spp. and
© 2023 Legba, Dougnon, Koudokpon, Mero,
Elovainio, Parry, Bankole and Haukka. This is an Salmonella enterica serovar Typhi while the most common gram-positives were
open-access article distributed under the terms Enterococcus spp. and Staphylococcus aureus. None of the laboratories listed
of the Creative Commons Attribution License Klebsiella pneumoniae among the three most common bacteria isolated from blood
(CC BY). The use, distribution or reproduction
in other forums is permitted, provided the cultures, although other evidence indicates that it is the most common cause of sepsis
original author(s) and the copyright owner(s) in Benin. Due to limited testing capacity, physicians most commonly use empirical
are credited and that the original publication in antibiotic therapy.
this journal is cited, in accordance with
accepted academic practice. No use, Conclusions: More resources are needed to develop laboratory testing capacity,
distribution or reproduction is permitted which technical skills in bacterial identification, AST, quality assurance, and communication
does not comply with these terms.
of results must be strengthened.

KEYWORDS

sepsis, blood culture, antibiotic susceptibility testing (AST), antibiotic prescribing, Benin

Introduction
Microorganisms entering the bloodstream can trigger sepsis, which is the body’s generalized
response to an infection and a life-threatening condition. Sepsis is the third most common cause
of death for children under the age of five (1). It is the most common cause of hospital deaths and
the leading cause of neonatal mortality, particularly in low- and lower-middle-income countries

Frontiers in Public Health 01 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

(LMICs) (2–4). Its incidence depends on complex interplay between Methods

factors related to the host, pathogen and health system response
(5). Several chronic diseases, sociodemographic factors, poor access Study setting, sample size and inclusion
to health care systems and quality of care are associated with the criteria
occurrence of sepsis and its case fatality rate (6).
Sepsis is most commonly caused by aerobic bacteria (7). In The study was conducted in Benin in West Africa. We wanted
LMICs the prevalence data are notably limited by the restricted ability to involve all laboratories from different parts of the country
to culture and identify organisms using standard microbiological with a functional bacteriology unit, therefore we contacted all
techniques. In addition to the pathogenicity of bacterial strains, the potential health centers and other stakeholders, such as the
the major concern is their increasingly common resistance to the Association of Medical Biologists of Benin. There are almost 150
antibiotics used in the treatment of sepsis (8). For example, according authorized clinical diagnostic laboratories in Benin. An initial survey
to the World Health Organization (WHO), resistance of Klebsiella identified all those that conduct bacteriological tests, including
pneumoniae, the major cause of bloodstream infections, to the blood culture and/or AST and we subsequently identified 27 such
carbapenem antibiotics used as last-resort treatment has spread to all authorized laboratories. Two laboratories failed to respond to our
regions of the world (9). Antimicrobial resistance (AMR) is in fact request; thus, the study was carried out with 25 laboratories. The
one of the major challenges in the management of sepsis, in particular laboratories were at different levels of the healthcare system: 2
in LMICs (4, 10). from the central or national level, 4 from the intermediate or
The bacteriology laboratory has twofold strategic role in the departmental level (Departmental Hospital Centers); and 19 from
diagnosis of sepsis. Firstly, at the individual level, laboratory tests the peripheral level (Zone hospitals (HZs), health centers, unattached
such as a blood culture and antibiotic susceptibility testing (AST) pharmacy dispensaries (independent pharmacies, i.e., not integrated
confirm the clinical diagnosis by identifying the causative organism or connected to a hospital or health center), unattached maternity
and providing data on the susceptibility of the organism to antibiotics wards (health facilities providing only a maternity services, i.e., the
(11). Secondly, for the clinic, laboratory diagnostics provide relevant maternity unit is not integrated or connected to a hospital or health
local information as a basis for the empirical use of antibiotics. center and faith-based health centers).
Through both mechanisms, the test results contribute toward All but one laboratory were affiliated with a healthcare facility.
prescribing the appropriate antibiotic for the effective treatment of Ten facilities were private, 9 public and 6 faith-based hospitals located
sepsis. Optimal testing activity and utilization of results requires close in different parts of the country. At each site, we aimed to interview
collaboration between laboratory staff, physicians and pharmacists, a laboratory technician, a laboratory manager, three physicians and
who provide antibiotics to the patients based on the prescriptions by two pharmacists. We interviewed technicians who conduct AST
physicians (12). and/or blood culture on a daily basis. The laboratory managers were
In Benin, laboratory diagnostics of clinical conditions such in charge of the laboratory. The physicians and pharmacists were
as sepsis is very limited. As in sub-Saharan Africa in general, practicing, not administrative, staff. The physicians were working in
clinical laboratories are typically poorly linked to clinical services, the general medicine service, intensive care, pediatrics or emergency
insufficiently resourced, and, therefore, under-utilized (13, 14). medicine. Physicians sending patients to the non-hospital connected
Furthermore, access to laboratory tests and drugs depends on private laboratory were identified in the nearby hospitals for the
capacity of the patients to pay for them, which often leads to tests interview. Several of the healthcare facilities did not have an in-house
not being done, and appropriate antibiotics not purchased (15). There pharmacy or the in-house pharmacy had only one pharmacist. In
is no national surveillance data for sepsis in Benin. However, in a these cases, pharmacists from pharmacies located close to the facility
recent study in central Benin, antibiotic resistant pathogens such as or a pharmacy most frequently visited by patients were included in
K. pneumoniae, Salmonella enterica serovar Typhi and Staphylococcus the study. These private pharmacies were identified based on the
aureus were isolated from blood cultures (14). Studies on other recommendation of the interviewed physicians.
clinically important pathogenic bacteria show AMR to be common In total, we interviewed 25 laboratory technicians, 25 laboratory
in Benin (16, 17). managers, 62 physicians and 47 pharmacists. Before data collection,
We wanted to find out the number and performance level informed consent was obtained from all participants. The interviews
of Beninese clinical laboratories that conduct microbiological tests were conducted anonymously.
required for the management of sepsis, especially blood cultures
and AST, across the country. We also studied whether physicians
make sufficient use of laboratory results when prescribing antibiotics
and whether pharmacists have sufficient knowledge when dispensing Data collection and analysis
antibiotics when there is a limited selection. We therefore conducted
a qualitative study to investigate understanding, perceptions, The data was collected between June 7 and June 19, 2021. We
knowledge, skills and practices of these three groups of professionals. used structured on-site, face-to-face interviews using tablets with
We also studied resources available to them and their professional a KoBoToolbox application (www.kobotoolbox.org/). The digitized
development needs. We conducted face-to-face interviews using data collection allowed for online data transfer and real-time quality
structured questionnaires with all three groups and supplemented control by a core team in Cotonou checking all forms and correcting
the results with discussions and visits to selected clinical laboratories. the missing or inaccurate information immediately.
The results will be used to develop appropriate policies for antibiotic The data collected included general information about the
prescription and stewardship, to improve sepsis management and hospital/laboratory and on the participating professionals.
to strengthen diagnostics and treatment of infectious diseases From the laboratory staff, we collected data on the practices
in general. related to blood cultures and AST. The questions were largely

Frontiers in Public Health 02 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

FIGURE 1
Distribution of the heathcare facilities with laboratories conducting blood cultures and/or antibiotic susceptibility testing (AST) in Benin. CNHPP, Centre
National Hospitalier et Universitaire de Pneumo-phtisiologie; CNHU, Centre National Hospitalier Universitaire; CDU, Centre de Diagnostics et d’Urgences;
CHD, Centre Hospitalier Départemental; HZ, Hôpital de Zone.

based on the WHO standard procedures (18). Physicians were Results

interviewed about their practices of requesting a blood culture
and AST before prescribing antibiotics and the interpretation Study sites and background information
of AST results. Pharmacists were asked about their knowledge
and practice in dispensing antibiotics for the treatment All 25 clinical bacteriology laboratories participating in the study
of sepsis. conducted AST but only 6 of them also blood cultures (Figure 1).
Data analysis was performed using SPSS version 24. The analysis We inquired about the availability of certain basic resources of the
was essentially descriptive and qualitative. A content analysis was laboratories to assess the general possibility of laboratories to do
carried out regarding the goals related to the identification of clinical microbiology work. Several laboratories lacked such basic
capacity development needs aimed at improving the blood culture resources as a microbiological safety cabinet, a freezer and internet
and AST practices. connection (Supplementary Table S1). The background information

Frontiers in Public Health 03 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 1 Practices related to a blood culture in the six laboratories that

of the interviewed staff showed that they all had a professional degree
provide the test.
(Supplementary Table S2).
Parameters Number of laboratories
(N = 6)
Blood culture practice Person taking the blood culture sample
Nurse 6
The key data obtained from the 5 health care facilities and
Technician 4
1 independent laboratory that conduct blood cultures are shown
in Table 1. In all the laboratories, the technicians indicated that Medical doctor 3
they follow a standard blood sample processing procedure. They, Trainee (Intern) 2
for example, checked that the blood sample was taken correctly,
Not known (external samples) 3
appropriate bottles were used (aerobic blood culture bottles for
adults or children), volume of blood was sufficient, and the weight Blood culture vials used
in relation to the recommended average was acceptable. The Biomerieux 4
recommended incubation temperature (35–37◦ C) was respected by
Liofilchem 2
all and the bottles grown in ordinary incubators were checked once,
twice or more often per day for growth. Two laboratories used the Average monthly number of blood cultures
BacT/ALERT automated system (BioMérieux, France), which signals Between 2 and 10 3
when there is growth in the bottle. In the case of a positive blood Between 11 and 20 1
culture, Gram staining was used to verify bacterial growth. The
20 or more 2
preliminary results were communicated to physicians to assist in
early treatment. Verification of proper blood sampling
Where growth in a blood culture bottle was detected, an aliquot Use of appropriate blood collection vials 6
was cultured on solid growth media (Table 1). For testing the quality
Volume of blood collected 5
of media, all the laboratories conducted a sterility test by incubating
the media plates at 35–37◦ C for 24 h. However, only 4 conducted the Attached patient information 4

performance test by growing some reference strains on the plates at Weighting the blood collection vials 1
35–37◦ C for 24 h. According to the interviews, the most common
Incubation of blood culture bottles in
bacteria isolated from blood cultures were: Gram-negative bacteria
Escherichia coli, Salmonella spp., Salmonella enterica serovar Typhi Ordinary microbiological incubator 4

and Proteus mirabilis, and gram-positive bacteria Enterococcus spp. Biomerieux automated system 2
and Staphylococcus aureus (Table 2). Incubation time of blood culture bottles
Three laboratories out of the six sometimes received samples
More than 5 days 4
from other hospitals. Yet, half of the laboratories processed a
maximum of 10 blood samples a month. Limiting factors mentioned 4 days 1
by the technicians and the laboratory managers for the low sample 1 day 1
numbers were: availability and cost of the culture media (all
Incubation temperature of blood culture bottles
laboratories used commercial blood culture vials) and other reagents,
the method used, the lack of an automated system, the limited level 35–37◦ C 6
of competence of the laboratory personnel and physicians’ lack of Incubator temperature control procedure
knowledge about the importance of early antibiotic therapy.
Built-in thermometer 5
When asked about their needs concerning conducting blood
cultures, the staff expressed their need for equipment and Separate thermometer in the incubator 1

consumables such as an automated blood culture machine and Frequency of checking the bottles in the incubator for growth
blood culture bottles. The laboratories also desired continuous
Once a day 1
training of staff to reinforce and develop their skills in detection
Twice a day 2
of difficult-to-cultivate bacteria, standardization of procedures and
conducting AST. More than twice a day 1

Machine signals in case of growth 2

Indication for a positive blood culture

Antibiotic susceptibility testing practice Uniform or subsurface turbidity 4

Gas production 3
Most of the laboratories used the disc diffusion method for AST
(Table 3). They mainly followed the Comité de l’antibiogramme de Signal of the machine 2

la Société Française de Microbiologie (CA-SFM) or the European Hemolysis 2

Committee on Antimicrobial Susceptibility Testing (EUCAST)
Flocculent deposit on the blood layer 2
Standards. Questions in our questionnaire were designed to assess
(Continued)
the compliancy with the standard and the quality of work. For

Frontiers in Public Health 04 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 1 (Continued)
Although all the laboratories included in the study conducted
AST, many of them did notably few tests, even <10 per month
Parameters Number of laboratories
(N = 6) (Table 3). The factors that limited the number of tests included (i)
limited demand; (ii) insufficient human resources; (iii) insufficiency
White grains on surface or deep in the 1
neck or shortage of laboratory equipment and consumables; (iv) negative
blood culture samples; and (v) prescription of antibiotics by the
Culturing after the predetermined time 1
physicians without requesting for a laboratory test. Staff in 17 of
Coagulation of the broth 0 the 25 laboratories expressed their need to have their capacity
A surface film 0 strengthened to conduct quality AST. Specifically, capacity building
in bacterial identification techniques, interpretation of AST results,
Staining method for positive blood culture
quality control and choosing antibiotic discs were mentioned.
Gram stain 6

Agar plates used for positive blood cultures

Ready to use agar plates 2 Physicians’ prescription practice and
Agar plates prepared in laboratory 2 utilization of laboratory results
Both 2
To get a general understanding of the use of antibiotics to
Growth media for positive blood cultures treat infections, we asked the physicians for all the reasons for
Blood agar 6 prescribing antibiotics (Supplementary Table S4). Digestive tract
Chapman (biochemical identification 5
infections (85.5%) were the most common infections treated
media) with antibiotics, followed by ear, nose and throat infections and
upper respiratory infections (82.3%). Only one case of sepsis
MacConkey 5
was mentioned, probably because sepsis was not considered as
Chocolate 5
primary diagnosis in most cases. Only 50% of the respondents
Eosin methylene blue agar 3 stated that they used a protocol defined by their hospital for the
Sabouraud agar 2
prescription of antibiotics. Among the physicians, who did not
use any protocol, 83.9% justified it by the hospital’s lack of a
Simmons citrate agar (biochemical 1
identification media)
protocol for antibiotic therapy. In cases where antibiotics were
used as a first-line treatment without bacteriological testing results,
Bile esculin agar (biochemical 1
the antibiotics most commonly prescribed were penicillins, such as
identification media)
penicillin G, amoxicillin and ampicillin, in 87.1% of cases, followed
Control procedure for media by cephalosporins, such as cefoxitin, ceftriaxone and ceftazidime, in
Sterility test 6 82.3% of cases (Supplementary Table S4).
Performance test 4
In treating sepsis, nearly all of the physicians interviewed
recognized that the identification of the pathogen should influence
Visual observation 1
the antibiotic therapy. However, in practice, empirical treatment was
Conservation of blood culture isolates used in most cases due to the lack of testing capacity (Table 4).
Yes 5 Physicians’ knowledge of various factors responsible for antibiotic
resistance was inadequate (Table 4). 80% of them recognized the
Communication of blood culture results to a physician
inappropriate choice of antibiotics as one of the probable causes
In person 4 of therapeutic failures in curing the bacterial infections. Prescribed
Phone 4 antibiotics not purchased by the patient, non-compliance with
treatment, superinfections, resistant bacteria and low product
Multiple ways 3
efficiency were also mentioned.
Electronic transmission 1 Of the 62 physicians interviewed, 54.8% had received 1 to 5
patients suspected to have sepsis in the preceding 12 months, while
9.7% had received none (Table 5). The patients were of all ages,
although most commonly children. According to the physicians,
example, Muller-Hinton culture medium poured into petri dishes
in determining the diagnosis of sepsis, clinical symptoms were
must be 4 ± 0.5 mm thick as defined by the CA-SFM standard. This
most important (91.1%) followed by blood culture results (73.2%).
measure was respected by 23 of 25 laboratories. Only 21 laboratories
Third generation cephalosporins, ceftriaxone (82.1%) and cefotaxime
used McFarland standards to determine the concentration of the
(46.4%), were the most commonly used antibiotics in the treatment
inoculum, while the rest assessed it by eye. The antibiotics for AST
of sepsis.
were primarily (72%) selected on the basis of CA-SFM/EUCAST
standards (Table 3). The other criteria mentioned for selecting the
antibiotics to be tested were: identity of the strain to be tested (64%)
and availability of antibiotics in the laboratory (36%). Only 18 (72%) Dispensing of antibiotics by pharmacies
of the laboratories performed quality control of the disks. The reasons
given for not performing quality control were: lack of materials, When the pharmacists were asked about antibiotics they provide
reference strains and expertise. for sepsis, they answered that the medical prescription does not

Frontiers in Public Health 05 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 2 Most commonly isolated micro-organisms from the blood cultures in the six laboratories culturing blood.

Laboratory Most commonly isolated Second most commonly Third most commonly
micro-organism isolated micro-organism isolated micro-organism
1 Salmonella spp. P. mirabilis S. aureus

2 Proteus spp. E. coli S. aureus

3 E. coli Salmonella Typhi S. aureus

4 E. coli Staphylococcus spp. Enterococcus spp.

5 S. aureus E. coli Streptococcus pneumoniae

6 Yeast S. aureus E. coli

include the indication and therefore they do not know which that the lack of equipment and consumables for blood cultures
antibiotics are for sepsis. However, they made some general leads the hospital to transfer patients to the hospitals in Djougou
suggestions for improving antibiotic delivery practices for sepsis or Tanguiéta, which are located in a distance of about an hour’s
treatment. These were (i) including all the required information drive. The latter hospitals are supported by the Catholic Church
in the prescription, (ii) having only medical doctors prescribe (Order of Malta Hospital in Djougou and St Jean de Dieu Hospital
antibiotics, (iii) having pharmacists check the correctness of the in Tanguieta) and have better laboratories than the governmental
prescription before dispensing the antibiotic, (iv) better control hospitals, due to better funding. In general, the public hospitals are
mechanism for dispensing antibiotics on medical prescription, and seriously underfunded, and consequently most of the laboratories
(v) general awareness raising and training on antibiotics, especially conducting blood cultures and AST are either private or faith-based,
for rural population. as shown by our study.
Besides the poor access to the bacteriological diagnostics, the
quality of laboratory results is a problem. In our questionnaires,
Discussion we had many questions related to the quality of testing and quality
control practices, since erroneous results can lead to inappropriate
All 25 laboratories covered by this study conducted AST, but only treatment of a patient. The laboratory staff indicated that they follow a
six conducted blood cultures. This illustrates the limited capacity for standard in processing blood samples in the laboratory. For example,
microbiological diagnostics of sepsis in Benin. In the conduct of a they controlled the volume of blood collected, since the sensitivity
blood culture, automated incubation and growth monitoring devices of blood culture depends on the volume. Blood culture bottles were
have almost become a standard in high-income countries, whereas incubated typically for 5 days and checked daily for bacterial growth.
this is far from being the case in LMICs (19). Most of the laboratories In literature, some authors recommend blind sub-culturing within
involved in our study had only very basic microbiological equipment, the first 24 h of incubation as an effective strategy for rapid detection
but two laboratories used an automated system for blood culturing. but the recent study in Benin did not recommend it because of
This equipment was provided by foreign partners rather than by increased work load and risk of contamination (14). Regardless of the
the Beninese government. However, not even these laboratories good intentions of the laboratory staff, our survey showed that there
conducted anaerobic blood cultures. In general, the main reasons is discrepancy between the standard procedure and daily practice.
reported by our interviewees for not conducting any blood cultures The major deficiency reported by the laboratory staff themselves
were the lack of equipment and the high cost. The situation is similar was related to identification of bacteria in case of a positive
in most LMICs which face many challenges in implementing blood blood culture. They reported Salmonella spp., E. coli, S. enterica
cultures due to financial, logistical and infrastructural constraints serovar Typhi, Enterococcus spp. and Staphylococcus aureus to be
(19). As a detail, it is worth mentioning that only 40% of the the most commonly isolated bacteria from blood cultures. These
laboratories surveyed had a microbiological safety cabinet. This findings partly match the results of the recent study in Benin,
compromises the safety of the staff as well as the quality of work (20). indicating Klebsiella pneumoniae, S. enterica serovar Typhi, S.
Even in the six laboratories conducting blood cultures, the aureus, E. coli, Enterobacter cloaceae and non-typhoidal Salmonella
number of samples processed was very low. Furthermore, the spp. as the most common isolates from sepsis in a Boko district
availability of microbiological testing in Benin is geographically hospital in central Benin (14). In another study, the most common
very biased. In 8 out of the 12 departments there is no laboratory bacteria isolated from neonatal sepsis cases in Africa were K.
conducting a blood culture, none in the whole central Benin. For pneumoniae, Klebsiella michiganensis, S. aureus, Serratia marcescens
patients this means, according to our discussions with the laboratory and Burkholderia cepacia (4). However, the laboratories interviewed
staff, that an accompanying person is obliged to travel several did not seem to be able to identify Klebsiella consistently, not
hundred kilometers to collect blood culture bottles from Cotonou, even in the hospital involved in the study of Ombelet et al. (14),
have the patient’s sample taken in a treating hospital and take since none of them mentioned Klebsiella among the three most
the bottle back to Cotonou, where two of the main laboratories common isolates from the blood cultures. This is possibly due to the
conducting blood cultures are located. They might have to make the difficulty in distinguishing between Klebsiella and E. coli. Also earlier
same round trip to get the results. In the north-west, the healthcare observations on identification of K. pneumoniae have indicated
facilities without possibility for blood culture may forward the patient limited accuracy in many LMICs (21). However, K. pneumoniae
to another facility. For example, at the Atacora CHD, they reported might indeed be one of the most common causes of sepsis in

Frontiers in Public Health 06 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 3 Practices related to antibiotic susceptibility testing (AST) in the 25 TABLE 3 (Continued)
laboratories that provide the test.
Parameters Proportion of laboratories
Parameters Proportion of laboratories (N = 25)
(N = 25)
n %
n %
1h 1 4
Average monthly number of samples tested for antibiotic
susceptibility 0 1 4

≤10 6 24 Incubation time of the agar plates

11–20 6 24 18–24 h 25 100

>20 13 52 Tools used for measuring the diameters of the inhibition zones
Ruler 18 72
AST method available
Vernier caliper 5 20
Antibiotic disc diffusion 22 88
Ruler and caliper 2 8
E-test 3 12
Standards for interpretation of the AST results
Agar medium used for AST
CA-SFM/EUCAST 15 60
Muller Hinton 24 96
CA-SFM/EUCAST and the disc 5 20
Cled 1 4 manufacturer’s instructions
Eosin Methylene Blue 1 4 Disc manufacturer’s instructions 5 20
Thickness of the medium in the petri dish Storage temperature of the antibiotic discs
4 mm 23 92 Refrigerator 24 96

>4 mm 2 8 Ambient room temperature 3 12

Procedure for determining inoculum Freezer 1 4

By eye without a standard 21 84 Carrying out quality control of antibiotic discs

By eye using Mc Farland standard 4 16 Yes 18 72

Inoculum spreading techniques Procedure for the disc quality control

Swabbing 18 72 Use of a reference bacterial strain 13 52

Flooding 10 40 Use of bacterial reference strain and 3 12

R
determination of MIC
Automated Biomerieux VITEK 3 12
Compact system Determination of MIC 2 8

Criteria for selecting the antibiotics Reasons for not performing quality control

CA-SFM/EUCAST 18 72 Lack of materials 3 12

Depends on the bacterium to be tested 16 64 No reference strains 2 8

Depends on availability of antibiotic 9 36 Lack of know-how 2 8

discs in laboratory
Skills improvement needed
Depends on availability of antibiotics in 1 4
pharmacy Yes 17 68

CLSI 1 4 No 8 32
CA-SFM, Comité d’Antibiogramme de la Société Française de Microbiologie; EUCAST,
Number of antibiotic discs tested per strain European Committee on antimicrobial susceptibility testing; CLSI, clinical and laboratory
≤10 7 28 standards institute; MIC, minimum inhibitory concentration.

11–20 17 68

>20 1 4

Procedure for depositing antibiotic discs on agar plates

Africa, with potentially high virulence and multidrug resistance
Sterile forceps 23 92 properties (22). Therefore, its epidemiological surveillance should be
Antibiotic disc dispenser 7 27 a priority. One of the laboratories included in our study reported
yeast as the most commonly isolated microorganism from blood
Time from depositing of antibiotic discs to placing the agar
plates into incubator cultures. Indeed, yeasts are among the microorganisms isolated from
bloodstream infections, but they are a relatively rare finding (4,
20 min 12 48
14, 23, 24). It can therefore be assumed that, given the very low
10 min 11 44 number of positive blood cultures in the laboratory in question,
(Continued) contamination and misidentification distorted the results. All the

Frontiers in Public Health 07 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 4 Physicians’ knowledge and practices in prescribing antibiotics. TABLE 4 (Continued)

Knowledge and practices Proportion of physicians Knowledge and practices Proportion of physicians
(N = 62) (N = 62)
n % n %
Knowledge of the concept of antibiotic resistance Financial capacity of the patient 1 1.6

Knowledge of natural resistance Probable reasons for encountered antibiotic treatment failures

Yes 56 90.3 Inappropriate choice of an antibiotic 42 79.2

Means of recognition of antibiotic resistance Non-compliance of the patient with 38 71.7

treatment
Treatment failure and persistence of 35 57.4
symptoms after normal duration of treatment Antibiotic not purchased by the patient 23 43.4

Results of AST 26 42.6 Superinfections 19 35.8

Factors believed to be responsible for antibiotic resistance Resistant pathogens 4 7.5

Poor quality of antibiotics available 55 88.7 Low antibiotic efficiency 1 1.6

Self-medication 54 87.1 Consulting an infectious disease specialist for the therapeutic

protocol
Misuse of antibiotics 42 67.7
Yes 5 8.1
Non-compliance with the dosage 37 59.7
Reason for not consulting an infectious disease specialist
Inadequate duration of the antibiotic therapy 26 41.9
No infectiologist in the health facility 45 80.4
Criteria other than AST for the choice of antibiotics
Has the necessary knowledge 6 10.7
General knowledge about effective antibiotics 51 82.3
Postgraduate training for antibiotic prescribing
Availability of the antibiotic in Benin 43 69.4
Yes 16 25.8
Previous experience with the effectiveness of 34 54.8
the antibiotic

Purchasing power of the patient 31 50

Availability of the antibiotic in the region 17 27.4 technicians we interviewed were particularly interested in improving
Cost of the antibiotic 7 11.3 their knowledge and practices in bacterial identification methods.
Our results showed that the physicians recognized the
Recommendation of a colleague 4 6.5
importance of a blood culture in the diagnosis of sepsis, but in
Prescription practices practice, the rate of testing was very low. There appears to be a
Identification of the bacterium influences antibiotic therapy vicious circle, where physicians do not request laboratory tests due
Yes 60 96.8
to their non-availability, patient’s inability to pay for them, slow
processing and unreliability of the results, and the low demand
Important criteria for the choice of antibiotics
for laboratory tests leads to poor resourcing of laboratories.
Results of AST 55 88.7 Consequently, physicians prescribe antibiotics on a probabilistic
Experience with the effectiveness of an 39 62.9 basis. Furthermore, hospitals have no standardized guidelines to
antibiotic support prescription practice. Only 5 of the physicians interviewed
Usually prescribed antibiotics 29 46.8 reported consulting infectious diseases specialists when prescribing
antibiotics. 45 specifically mentioned the absence of infectious
Bacterial species and infection site 1 1.6
disease specialists from their health facility.
Clinical condition of the patient 1 1.6 According to our survey, factors limiting testing included
Preference for generic or brand-name antibiotics insufficient properly trained personnel, insufficient or broken
laboratory equipment and lack of consumables. The analysis of
Generic and brand-name antibiotics 27 43.5
practices related to AST revealed technical deficiencies concerning
Brand-name antibiotics 23 37.1 the choice of antibiotic discs and the quality control of the discs,
Generic antibiotics 18 29 media and growth of reference bacteria. Only 18 (72%) of the
laboratories performed quality control of the discs although it is
Factors that influence the preference to prescribe an antibiotic
strongly recommend by the EUCAST standards to ensure that
Cost of the product 42 67.7 efficacy has not been recuded by e.g., poor storage or other
Unavailability of brand-name antibiotics 38 61.3 conditions. The antibiotics chosen for AST were sometimes chosen
Quality of generics 33 53.2
based on availability of antibiotics in the pharmacies near the
hospital. This might be rational considering the treatment but does
Prescription protocol of the healthcare center 13 21
not provide proper surveillance information for the local situation.
Effectiveness of an antibiotic 2 6.5 The majority of technicians expressed their need and willingness
(Continued) for further training in AST. It was also seen to be important to

Frontiers in Public Health 08 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

TABLE 5 Management of sepsis and types of antibiotics prescribed.

pharmacists indicated as a major problem due to the insufficient
training of nurses. Thus, in future studies exploring the prescription
Parameters Proportion of physicians
(N = 62) of antibiotics and in training programs, it would be important to
include nurses as well as physicians. Furthermore, in West Africa,
n %
many antibiotics are sold in pharmacies without prescription (only
Number of patients with suspected sepsis, encountered one pharmacy in our study admitted this) and by street vendors
during the last 12 months
(26). The different galenic forms can also lead to confusion among
1–5 34 54.8 prescribers and patients in the correct use of the antibiotics (27).
>5 22 35.5 Therefore, optimizing the monitoring of antibiotic delivery is also a
way to improve antibiotic use practices (28). Moreover, the quality of
None 6 9.7
antibiotics and many other medicines in LMICs is often substandard
Profile of patients with sepsis (27). One factor that contributes to a partial or total reduction in
People 15 years of age or older 26 46.4 the quality of antibiotics in the hot and humid climate in West
Children under 5 years of age 24 42.9
Africa is poor storage (29). This issue was also mentioned in our
interviews. Proper storage of antibiotics is costly and requires well
Children 5–14 years of age 22 39.3
trained personnel, which is in short supply in West African countries.
Newborns 19 33.9 Although it is well-known that poor quality of antibiotics leads to an
Pregnant women 9 16.1 increase in multi-resistant bacteria and the risk of therapeutic failure,
very few LMICs have a quality control agency to monitor the quality
Patients with comorbidities 8 14.3
of medicines (30, 31).
Method of diagnosing sepsis in a patient Our main reason for undertaking this study was to gain
Clinical symptoms 51 91.1 understanding of the base-line level of the bacteriological laboratory
diagnostics of blood culture and AST for bacteria that cause sepsis.
Blood culture 41 73.2
We utilized the results in designing a training module for laboratory
Catecholamine testing 1 1.8 technicians to improve their competence (details will be reported
Blood count 1 1.8 elsewhere). The laboratory staff that participated in the study
Means of communication with the laboratory in case of blood
appreciated our effort to contact all laboratories across the country
culture and AST were requested and address their concerns. We are planning a follow-up training
course concentrating on identification of key bacteria causing sepsis.
By phone 54 96.4
Since successful treatment and prevention of infections requires
In person 23 41.1 multi-professional collaboration, we also interviewed physicians and
By an intermediary person (caregiver, 13 23.2 pharmacists on their knowledge of sepsis and usage of antibiotics.
nurse, etc.) Based on the obtained results, we have organized events to bring
By e-mail 1 1.8 the different professional groups and the national health authorities
Types of antibiotics used in treatment of sepsis
together to discuss the best practices in the local settings.
In conclusion, we recognize an urgent need to increase the
Ceftriaxone 46 82.1
availability and quality of blood cultures and AST for improved
Cefotaxime 26 46.4 sepsis management throughout Benin. Laboratories with a clinical
Ceftazidime 19 33.9 bacteriology unit must be provided with appropriate equipment and
more consumables to ensure that there is at least one laboratory able
Imipenem/cilastatin 18 32.1
to conduct necessary diagnostics in each of Benin’s 12 departments.
Meropenem 14 25 The laboratory staff involved in this study themselves expressed their
Cefepime 10 17.9 need and willingness to strengthen their skills in conducting both
blood cultures and AST. Also reinforced collaboration between the
Ampicillin and sulbactam 8 14.3
laboratories, physicians and pharmacists is necessary for improved
Levofloxacin 7 12.5 sepsis management.
Gentamycin 7 12.5

Clindamycin 6 10.7
Data availability statement
Ciprofloxacin 3 5.4

Metronidazole 2 3.6 The original contributions presented in the study are included in
the article/Supplementary material, further inquiries can be directed
Piperacillin and tazobactam 1 1.8
to the corresponding author.

standardize the practices at the national level to improve the quality Author contributions
of microbiological testing.
In Benin it is a common, but unofficial, practice that nurses BL, VD, HK, SM, RE, MP, HB, and KH designed the study. BL,
rather than medical doctors prescribe antibiotics (25), which the VD, and HK collected and analyzed the data. BL wrote and VD,

Frontiers in Public Health 09 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

HK, SM, RE, MP, HB, and KH revised the manuscript. All authors Conflict of interest
contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the
absence of any commercial or financial relationships that could be
Funding construed as a potential conflict of interest.

The study was funded by the grant from the Finnish Ministry for
Foreign Affairs to the Physicians for Social Responsibility (PSR)— Publisher’s note
Finland for the Quality laboratory testing training project in Benin.
All claims expressed in this article are solely those of the
authors and do not necessarily represent those of their affiliated
Acknowledgments organizations, or those of the publisher, the editors and the reviewers.
Any product that may be evaluated in this article, or claim that may
We are grateful to Drs. Clement Ahoussinou and Mireille be made by its manufacturer, is not guaranteed or endorsed by the
Mintogbe who assisted us in the implementation of the survey. publisher.
We also thank Kevin Sintondji, Arielle Kounou, Alida Oussou,
and Beau-Gard Hougbenou, young researchers at the University of
Abomey-Calavi (Benin) for conducting the interviews; and the staff Supplementary material
of the hospitals involved and the interviewees for their willingness
to participate in the study. Finally, we are grateful to Heli Salmi, The Supplementary Material for this article can be found
Mari Laaksonen, and Anu Kantele from PSR-Finland for fruitful online at: https://www.frontiersin.org/articles/10.3389/fpubh.2022.
discussions on the project design. 1088590/full#supplementary-material

References
1. Prual A. Le nouveau-né en Afrique de l’Ouest et du Centre: comprendre pour agir. 14. Ombelet S, Kpossou G, Kotchare C, Agbobli E, Sogbo F, Massou F, et al.
Sante Publ. (2020) 1:7–15. doi: 10.3917/spub.200.0007 Blood culture surveillance in a secondary care hospital in Benin: epidemiology of
bloodstream infection pathogens and antimicrobial resistance. BMC Infect Dis. (2022)
2. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, Tsoi D, Kievlan 22:119. doi: 10.1186/s12879-022-07077-z
DR, et al. Global, regional, and national sepsis incidence and mortality,
1990–2017: analysis for the Global Burden of Disease Study. Lancet. (2020) 15. Koura KG, Garcia A, Todoégnon B, Deloron P, Cot M, Faucher J-F. Prevalence and
395:200–11. doi: 10.1016/S0140-6736(19)32989-7 factors related to antibiotic prescription in Benin: a school-based study. Acta Trop. (2013)
127:87–90. doi: 10.1016/j.actatropica.2013.04.004
3. Do SN, Luong CQ, Pham DT, Nguyen MH, Nguyen NT, Huynh DQ, et al. Factors
relating to mortality in septic patients in Vietnamese intensive care units from a subgroup 16. Dougnon V, Assogba P, Mohammed J, Agbankpe J, Déguenon E, Fabiyi K,
analysis of MOSAICS II study. Sci Rep. (2021) 11:18924. doi: 10.1038/s41598-021-98165-8 et al. Urinary tract infections in Benin: exploring the virulence factors and antibiotic
resistance and virulence genes among bacterial isolates. Int J Pathog Res. (2021) 7:28–
4. Milton R, Gillespie D, Dyer C, Taiyari K, Carvalho M, Thomson K, et al. Neonatal 36. doi: 10.9734/ijpr/2021/v7i130174
sepsis and mortality in low-income and middle-income countries from a facility-based
birth cohort: an international multisite prospective observational study. Lancet Glob 17. Adzavon P, Dougnon TV, Koudokpon H, Legba B, Agbankpe J, Fabiyi K, et al.
Health. (2022) 10:e661–72. doi: 10.1016/S2214-109X(22)00043-2 Resistance profile of Staphylococcus strains and detection of the Mec A, Van A
and Van B genes in private hospitals in Benin. Afr J Microbiol Res. (2022) 16:184–
5. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med. (2013) 9. doi: 10.5897/AJMR2021.9537
369:840–51. doi: 10.1056/NEJMra1208623
18. Leenstra T, Kooij K, Tambic A, Nahrgang S, van de Sande-Bruinsma N. Proof-
6. World Health Organization. État septique. Geneva: World Health Organization of-Principle Antimicrobial Resistance Routine Diagnostics Surveillance Project: Protocol.
(2020). Available online at: https://www.who.int/fr/news/item/08-09-2020-who-calls- Copenhagen: WHO Regional Office for Europe (2018). Available online at: https://www.
for-global-action-on-sepsis---cause-of-1-in-5-deaths-worldwide (accessed July 23, euro.who.int/__data/assets/pdf_file/0007/360529/WHO-PoP-project-eng.pdf (accessed
2022). 17 May 2022).
7. Gotts JE, Matthay MA. Sepsis: pathophysiology and clinical management. BMJ. 19. Ombelet S, Barbé B, Affolabi D, Ronat J-B, Lompo P, Lunguya O, et al. Best
(2016) 353:i1585. doi: 10.1136/bmj.i1585 practices of blood cultures in low- and middle-income countries. Front Med. (2019)
8. Ombelet S, Ronat J-B, Walsh T, Yansouni CP, Cox J, et al. Clinical bacteriology 6:131. doi: 10.3389/fmed.2019.00131
in low-resource settings: today’s solutions. Lancet Infect Dis. (2018) 18:e248– 20. Arfin Qasmi S, Pirzada S, Ghani A, Mohsin S. Survey on proper and
58. doi: 10.1016/S1473-3099(18)30093-8 safe use of biological safety cabinets (BSCs) in research, bio-medical and animal
9. World Health Organization. Antimicrobial Resistance. Geneva: World Health laboratories in Karachi, Pakistan a cross sectional study. Int J Biosaf. (2020) 2:77–
Organization (2021). Available online at: https://www.who.int/news-room/fact-sheets/ 80. doi: 10.1016/j.jobb.2020.10.003
detail/antimicrobial-resistance (accessed July 28, 2022). 21. Osman KM, Yousef AMM, Aly MM, Radwan MI. Salmonella spp. infection in
10. Martin C, Brun-Buisson C. Prise en charge initiale des états imported 1-day-old chicks, ducklings, and Turkey poults: a public health risk. Foodborne
septiques graves de l’adulte et de l’enfant. Ann Fr Anesth Reanim. (2006) Pathog Dis. (2010) 7:383–90. doi: 10.1089/fpd.2009.0358
26:53–73. doi: 10.1016/j.annfar.2006.10.021
22. Müller-Schulte E, Tuo MN, Akoua-Koffi C, Schaumburg F, Becker SL.
11. Jacobs J, Hardy L, Semret M, Lunguya O, Phe T, Affolabi D, et al. High prevalence of ESBL-producing K. pneumoniae in clinical samples from
Diagnostic bacteriology in district hospitals in Sub-Saharan Africa: at the central Côte d’Ivoire. Int J Infect Dis. (2019) 91:207–9. doi: 10.1016/j.ijid.2019.
forefront of the containment of antimicrobial resistance. Front Med. (2019) 11.024
6:205. doi: 10.3389/fmed.2019.00205
23. Chukwumeze F, Lenglet A, Olubiyo R, Oluyide B, Oloruntuyi G, Ariti C,
12. Bronkhorst E, Gous AGS, Schellack N. Practice guidelines for clinical et al. Multi-drug resistance and high mortality associated with community-acquired
pharmacists in middle to low income countries. Front Pharmacol. (2020) bloodstream infections in children in conflict-affected northwest Nigeria. Sci Rep. (2021)
11:978. doi: 10.3389/fphar.2020.00978 11:20814. doi: 10.1038/s41598-021-00149-1
13. Nkengasong JN, Yao K, Onyebujoh P. Laboratory medicine in low-income 24. Gaballah A, Shawky S, Amer A. Microbiological profiles of
and middle-income countries: progress and challenges. Lancet. (2018) 391:1873– neonatal sepsis in northern Egypt. Microbes Infect Dis. (2022) 3:645–
5. doi: 10.1016/S0140-6736(18)30308-8 56. doi: 10.21608/mid.2022.129600.1265

Frontiers in Public Health 10 frontiersin.org

Legba et al. 10.3389/fpubh.2022.1088590

25. Sariola S, Butcher A, Cañada JA, Aïkpé M, Compaore A. Closing the GAP in focus on developing countries. Life. (2021) 11:528. doi: 10.3390/life1106
antimicrobial resistance policy in Benin and Burkina Faso. mSystems. (2022) 7:e00150– 0528
e00122. doi: 10.1128/msystems.00150-22
29. Taylor R, Shakoor O, Behrens R, Everard M, Low AS, Wangboonskul J, et al.
26. Belachew SA, Hall L, Selvey LA. Non-prescription dispensing of antibiotic Pharmacopoeial quality of drugs supplied by Nigerian pharmacies. Lancet. (2001)
agents among community drug retail outlets in Sub-Saharan African countries: 357:1933–6. doi: 10.1016/S0140-6736(00)05065-0
a systematic review and meta-analysis. Antimicrob Resist Infect Control. (2021)
30. Nebot Giralt A, Schiavetti B, Meessen B, Pouget C, Caudron JM, Marchal
10:13. doi: 10.1186/s13756-020-00880-w
B, et al. Quality assurance of medicines supplied to low-income and middle-
27. Ouedraogo AS, Pierre HJ, Bañuls AL, Ouedraogo R, Godreuil S. Émergence et income countries: poor products in shiny boxes? BMJ Global Health. (2017)
diffusion de la résistance aux antibiotiques en Afrique de l’Ouest : facteurs favorisants et 2:e000172. doi: 10.1136/bmjgh-2016-000172
évaluation de la menace. Med Sante Trop. (2017) 27:147–54. doi: 10.1684/mst.2017.0678
31. Ravinetto R, Pinxten W, Rägo L. Quality of medicines in resource-
28. Godman B, Egwuenu A, Haque M, Malande OO, Schellack N, Kumar limited settings: need for ethical guidance. Glob Bioeth. (2018) 29:81–
S, et al. Strategies to improve antimicrobial utilization with a special 94. doi: 10.1080/11287462.2018.1522991

Frontiers in Public Health 11 frontiersin.org