Introduction
The growing burden of antibiotic resistance (AR) is a serious global public health problem, presenting a significant threat to the success of treatment, prevention, and control of infectious diseases [
1]. While AR poses potential long-term problems globally, currently, patients with AR have worse outcomes including prolonged hospital stays, increased healthcare costs, and increased morbidity and risks of mortality [
2]. These factors are worsened in low-and-middle-income countries (LMICs) where infrastructure is lacking and the availability of second- and third-line antibiotic therapies that may be more effective are not widely available [
3]. Control of AR requires surveillance to understand the magnitude and burden in LMICs, but this is challenging in many countries because of limited financial and human resources for health capacity and poor laboratory infrastructure for microbiological diagnosis [
1,
4]. Increasing antibiotic consumption fueled by improving economies in LMICs poses a threat to the control of AR [
5].
In addition to the risks from AR, hospitalized patients in LMICs face a greater risk of contracting hospital-acquired infections (HAIs). The growing burden of HAIs which are driven by poor infection prevention and control (IPC) practices are among the drivers of antibiotic resistance. Evidence indicates a tightly interwoven relationship between antibiotic resistance and HAIs [
6].
Multidrug-resistant pathogens are a common cause of HAIs and place a heavy toll on patients and their families by causing illness, potential disability, excess costs and sometimes death [
7,
8].HAIs in high-income countries have been estimated to affect 5 to 15% of hospitalized patients in regular wards and 50% or more of patients in intensive care units(ICU) [
9]. Although estimates from LMICs suggest that the HAI rate is at least 3 times higher than in the USA [
10,
11], however, studies are limited, particularly in Africa, where the burden of HAIs remains largely unknown or underestimated. The lack of surveillance in Africa is due to the complexities in diagnosis and the limited resources required for surveillance to guide interventions. Even though sub-Saharan African countries have committed to control antibiotic resistance in their countries [
12], due to a lack of resources and other pressing issues there is, as yet, little effort at scaling up activities on prevention and control of antibiotic resistance. In the WHO African region, only Ethiopia and South Africa had a national antimicrobial resistance plans in place and neither of the countries in this region has a national antimicrobial surveillance system [
13].Further compounding the problem is the limited availability of data on antibiotic resistance by a large number of African countries [
14]. In a systematic review of the burden of antimicrobial resistance in West Africa, Sierra Leone was singled out among other countries due to paucity of the AR data available [
15].
Sierra Leone has a National IPC Policy with eleven core components including surveillance and control of antimicrobial resistance and HAIs [
16], and a 5-year strategic plan on antimicrobial resistance for 2017–2021.Yet, there are few activities relating to the implementation of surveillance of AR and HAIs in the country. As there is no previously published study on AR in patients with clinical features of HAI in Sierra Leone, our study aimed to determine the bacterial pathogens and their antibiotic resistance profile among patients with features of these infections in an urban tertiary hospital in Sierra Leone.
Methods
Study design and setting
The study used a cross-sectional study design to collect data from 164 adult (≥18 years) inpatients in a 300-bed tertiary hospital in Freetown Sierra Leone.
Connaught Hospital is one of the hospitals within the University of Sierra Leone Teaching Hospitals Complex which was established to provide clinical services, medical training and research. The hospital has four medical and five surgical wards, two private wards and an intensive care unit (ICU). The hospital’s bed occupancy rate at the time of data collection was estimated at 70% during the study.
Participant selection
Participants were recruited sequentially over an 18-week period from February to June 2018. A total sample of 164 participants was conveniently recruited from the medical wards of Connaught Hospital. All hospitalized patients for 48 h or more aged 18 years or older with clinical features of catheter-associated urinary tract infection (CAUTI) and healthcare-associated pneumonia (HAP) with no evidence of an overt or incubating infection at admission or with new-onset symptoms of an infectious process (whether or not they have an underlying structural lung or kidney disease) were eligible for inclusion. None of the patients had ventilator-associated pneumonia (VAP) as the study was limited to the medical wards.
Exclusion criteria were those under 18, patients unable to produce appropriate specimen, and those who declined consent.
Clinical criteria for CAUTI and HAP
Catheter-associated UTI
Patient had an indwelling urinary catheter that had been in place for more than 2 consecutive days in an inpatient location with at least one of the following signs or symptoms: fever (> 38.0 °C), suprapubic tenderness, costovertebral angle pain or tenderness, urinary urgency, urinary frequency or dysuria [
17].
Healthcare-associated pneumonia
Fever (> 38 °C) with at least one of the following features: new-onset cough or worsening cough, dyspnea or tachypnea, rales or bronchial breath sounds and low SPO
2 (< 92%) [
17,
18].
Specimen and data collection
Two teams of doctors and nurses, trained on the assessment of clinical features of HAI collected data from patients’ files as recorded by the managing teams, and by patients/relatives/ward nurses’ interviews. Baseline information on demographic characteristics and clinical details were collected from patients who met the clinical criteria for HAI. After appropriate labeling of sterile containers, urine and sputum specimens were collected using standard operating procedures.
Although standard criteria were not used to assess sputum quality, the quality of all sputum samples collected was assured by adequately educating the patients and providing appropriate supervision of the sputum collection. As all respiratory samples were sputum collected without induction, sputum samples with poor quality were discarded.
All the patients who had urinary catheters in situ had their catheters changed and urine specimen collected from the new catheters before attaching new urine bags. Latex catheters are routinely used under aseptic conditions in this facility. Severe prostration, urinary retention or incontinence, and altered level of consciousness are some of the indications for catheterizations.
All specimens were immediately transported in a specimen container to the laboratory, situated about 5 km away. As the laboratory was not operational on public holidays and weekends, specimens and data were only collected by the research team on non-holiday week days. When there was a delay in transportation, samples were temporarily stored at 4 °C.
Laboratory materials and methods
Chromogenic agar (CHROM agar Orientation medium and CHROM agar, France) was used as a selective and differential medium, while brain heart infusion (BHI) agar (Qingdao Hope Biotechnology, China) was used for purification and identification. GN/GP cards and AST-GN09/AST-GP67 cards were used for identification and antibiotic susceptibility testing of gram-negative bacteria/gram-positive bacteria. Gram stain kit (Zhuhai Baso Biotechnology, China) was used for determining gram stain reaction and bacterial morphology of isolates.
Isolation and purification
Upon arrival at the laboratory, urine and sputum samples were streaked onto the chromogenic agar plate within 3 h and incubated aerobically at 37 °C for 18 to 24 h. Where a bacterial growth was observed on the chromogenic agar plate, a single bacterial colony was picked up and then streaked onto a BHI agar plate for purity and a Gram stain. In order to ensure that all isolates were pure, all isolates were cultured at least twice but there were no discordant results.
Identification and antibiotic susceptibility test
A VITEK 2 compact system (bioMérieux, France) was used for identification and antibiotic susceptibility testing of isolates from pure cultures.
A solution of bacteria in saline was prepared in polystyrene tubes (bioMérieux, France) to 0.5~0.63MacFarland turbidity using DensiCHEK Plus turbidimeter (bioMérieux, France). Antibiotic susceptibility test was conducted by adding 145 μl (for gram-negative bacteria) or 280 μl (for gram-positive bacteria) of suspension into a new polystyrene tube as per the manufacturer’s instructions.
The isolates suspensions were loaded on the VITEK 2 compact system and incubated overnight at 37 °C. All results of cultures and resistance testing were dispatched to the research team and service providers within 24 h. Appropriate advice on antibiotic selection was given to the providers.
Discussion
This is the first study to assess antibiotic resistance among patients with clinical features of healthcare-associated infections in a tertiary hospital in Sierra Leone.
The proportion of catheterized patients patients in our study with an infection is similar to other studies on HAIs in sub-Saharan Africa (58.0% vs 68.7%) [
8] and Asia (58.0% vs 47.7%) [
19]. The high rate of infections associated with catheterization in our study could be due to a lack of hospital protocols on aseptic procedures, including catheterization. Therefore, policies and training on catheterization, as well as, other aseptic procedures are needed to prevent HAIs in hospitals in Sierra Leone and other LMICs. Policies on patients’ safety initiatives and education of healthcare workers on urinary catheterization has been shown to be effective preventive strategies for catheter-associated urinary tract infections(CAUTI) [
20]. Moreover, similar to an Ethiopian study, underlying co morbidities predisposed patients to HAIs, HIV being the most common [
8]. Improving the care of patients with chronic diseases like HIV, stroke and diabetes mellitus, thereby preventing complications and progression to the advanced stage, will help in the reduction of HAIs in LMICs.
A proportion of patients higher than recorded in other studies in LMICs [
21,
22] and high-income countries(HICs) [
23] used at least one antibiotic prior to enrollment. The high rate of antibiotic use by this cohort could be a predisposing factor to the high levels of resistance observed in this study, although it could also be explained by the effort of service providers to treat the patients’ febrile conditions. As with other studies, the high burden of comorbidities especially the HIV/AIDS burden in this hospital [
8,
24,
25] will also have influenced the use of antibiotics by service providers. Although there were existing antibiotic guidelines in this hospital, their use could not provide adequate guidance to the treatment of HAP and CAUTI. Nonetheless, antibiotics used by patients in this study were mainly first or second line similar to findings in a recent study on antibiotic use in this hospital [
26].
Participants reported other clinical features such as cough, dyspnea, dysuria and frequency of passage of urine. Even though most of the samples cultured were urine specimens, more patients reported respiratory symptoms than urinary symptoms. This is not surprising as there were difficulties in getting proper respiratory samples from very ill patients.
A retrospective review of culture data in Ghana showed a prevalence of bacterial isolates higher than observed in our study(57.7% vs 16.6%) [
27]. However, heterogeneity between study populations (community vs healthcare) and the wider spectrum of the cultured specimen in the Ghanaian study prohibit direct comparison. Similar to a systematic review and meta-analysis on HAIs in developing countries, gram-negative bacillary infections in both urinary and respiratory tract infections are the most recorded bacterial isolates in our study [
10]. Again, other studies on antibiotic resistance in the African region noted similar spectrum of bacterial isolates found in our study [
14].
Unlike a systematic review on ESBL-producing organisms in Africa [
28], a high rate of ESBL-producing gram-negative bacteria was detected in this study (58.0% vs 22.8%). Nonetheless, carbapenem-resistance rate in
A. baumannii, P. aeruginosa, and
Enterobacteriaceae were lower than reported in other LMICs [
29,
30] and some high-income countries [
31]
. These patterns of antibiotic resistance in this setting could be explained by the high rate of use of third generation cephalosporins and a lack in exposure of admitted patients in this hospital to carbapenems [
26]. Moreover, the carbapenem resistance rate to
E. cloacae may reflect elevated minimum inhibitory concentrations (MICs) they sometimes display or may also reflect an error on the susceptibility instrument.
Compared to a systematic review on antibiotic resistance in West Africa, resistance of
E. coli and
K. pneumoniae to ampicillin and ampicillin-sulbactam was slightly higher in our study (ampicillin: 93% vs 81% for
E. coli and 93% vs 90.1% for
K. pneumoniae) [
15], although
K. pneumonia resistance to ampicillin and other penicillins is expected as most strains are intrinsically resistant to these agents [
32].
Similar proportions of
E. coli and
K. pneumoniae resistant isolates were recorded for trimethoprim-sulphamethoxazole, though slightly higher than in the West African systematic review [
15]. These trends may signify shared risk factors and similar weaknesses in policies on antibiotic resistance in the sub region, indicating urgent action by member states of the West African community.
All the S. aureus isolates were MRSA. But as there were only two isolates, we should be cautious about generalizing this to all S. aureus isolates in our patient population. Moreover, the two MRSA isolates demonstrated excellent susceptibility to vancomycin, linezolid, tigecycline, and rifampicin.
Our study had several limitations, including the small sample size and its restriction to a single, urban study site and regular medical wards, making the findings not readily generalizable. However, as the hospital is the main referral center in the country, there may not be marked variations in antibiotic resistance patterns from inpatients in other hospitals.
The isolation of polymicrobial organisms in some urine specimens could be from delayed transportation of a few specimens to the laboratory, which is a few kilometers away from the hospital, though this is unlikely to affect the rate of antibiotic resistance in this hospital as isolates were pure microorganisms. Convenience sampling and the type of media used in laboratory processing are additional limitations of the study although this ought not to affect the overall results as most healthcare-associated infections are caused by facultative anaerobes that are not typically fastidious.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.