Bacterial Nosocomial Infections and Antimicrobial Susceptibility Pattern among Patients Admitted at Hiwot Fana Specialized University Hospital, Eastern Ethiopia

Nosocomial infections remain a major cause of mortality and morbidity worldwide. Despite the highly specialized interventions and policies, the rate of infection is still high due to the emergence of antimicrobial-resistant bacteria. This study described the prevalence of bacterial nosocomial infections and antimicrobial susceptibility pattern of isolates among patients admitted at Hiwot Fana Specialized University Hospital, Eastern Ethiopia. A hospital-based cross-sectional study was conducted among 394 nosocomial infection-suspected patients from March 2017 to July 2017. Data were collected using a structured questionnaire. Specimens from the respective site of infections were collected and examined for the presence of pathogenic bacteria and their antimicrobial susceptibility using standard culture and serological tests. Data were summarized using descriptive statistics. The prevalence of culture-confirmed bacterial nosocomial infection was 6.9% (95% CI: 4.3–7.9). Staphylococcus aureus (18.5%) was the most common isolate followed by Escherichia coli (16.7%). S. aureus showed 80% resistance to chloramphenicol and erythromycin, and 70% to cephalexin and tetracycline, respectively. A methicillin-resistant S. aureus made up 88.9% of all S. aureus isolates. Pseudomonas aeruginosa showed 83.7% resistance to each of ceftazidime and cephalexin, and 66.7% to chloramphenicol. The most common multidrug-resistant isolates were P. aeruginosa (30.4%) and S. aureus (21.7%). The prevalence of nosocomial infections in this study was comparable with other findings; however, the high rates of antimicrobial resistant isolates represent a substantial threat to the patients, communities, health care providers, and modern medical practices. Bacterial nosocomial infection treatment should be supported by culture isolation and antimicrobial susceptibility testing.


Introduction
A nosocomial infection (NI) (also known as hospitalacquired infection) is a localized or a systemic infection resulting from an adverse reaction to infectious agents or its toxins that develops in 48 hours or more after admission and was not incubating on admission [1,2]. e most common type of NIs are urinary tract infections (usually catheter associated) (31%) [3] followed by surgical site infections (SSIs) (17%), primary bloodstream infections (BSIs) (usually associated with the use of an intravascular device) (14%), and pneumonia (usually ventilator associated) (13%) [3,4]. e main bacteria associated with NIs are S. aureus, coagulase-negative staphylococci (CoNS), Streptococcus pneumoniae, Escherichia coli, P. aeruginosa, Haemophilus influenzae, Klebsiella pneumoniae, Acinetobacter, and Enterococci [5,6]. e transmission within the hospital occurs through crosscontamination of the patients via the contaminated hands of health care staffs who come in frequent contact with patients or through contaminated objects [4,7]. e emergence of antimicrobial-resistant bacteria has become a public health problem, creating a new burden on modern medical care in hospitals [8,9]. P. aeruginosa, S. aureus, Enterococci spp., Klebsiella spp., and Enterobacter spp. are the major resistant pathogens of concern [10,11]. Particularly, methicillin-resistant S. aureus (MRSA) is known to cause considerable morbidity and mortality in hospitalized patients [12,13]. e consequence of the infection caused by resistant bacteria lies in their ability to not only alter the outcome of critically ill patients but also reduce the chances of the treatment, prolong the duration of the hospitalization, increase the cost of health care, and make the spread of infection easier and the prevention more difficult [14,15].
Knowledge of proper antimicrobial prescription policy of a particular setting is crucial to optimize the management and reduction of the rate of NIs; however, the investigation of causative agents and their antimicrobial susceptibility profile are an essential prerequisite [16,17]. In Ethiopia, some comprehensive studies were conducted on NIs [18][19][20]; none of these include most of the sources of NIs and have not determined the antimicrobial susceptibility pattern of the causative agents. Moreover, a study reported in one region is not necessarily reflecting the status of other regions.
is study was carried out to assess the prevalence and antimicrobial susceptibility pattern of bacteria causing NIs among patients admitted at Hiwot Fana Specialized University Hospital, Eastern Ethiopia.

Study Setting and Design.
A cross-sectional quantitative study involving bacteriological analysis was conducted at Hiwot Fana Specialized University Hospital, Harar, Eastern Ethiopia from March 2017 to July 2017. Harar is the political and administrative town of the Harari Regional State and is located at 525 km from Addis Ababa, Ethiopia. ere are six hospitals and eight health centers in this region. Hiwot Fana Specialized University Hospital provides health care services and serve as a referral hospital for eastern parts of our country. It has the largest client load with an average bed occupancy rate of 83% (sources: Hiwot Fana Specialized University Hospital Annual Report of 2016). e hospital consists of six major wards: Medical, Surgical, Obstetrics, Gynecology, Malnutrition, and Pediatric wards.

Study Population.
Patients admitted to the Medical, Surgical, Obstetrics, Gynecology, Malnutrition, and Pediatric wards for more than 48 hours and who had a clinical evidence of NIs were included in this study.

Sample Size and Sampling Technique.
A single population proportion formula was used to calculate a sample size, assuming 95% confidence level, 3% margin of error, 10% predicted nonresponse rate, and 10.3% prevalence of NIs [1]. e final sample size was determined to be 433. e study participants were selected consecutively until the required sample size fulfilled.

Data and Specimen Collection. Patients admitted in
Medical, Surgical, Obstetrics, Gynecology, Malnutrition, and Pediatric wards were followed prospectively for the development of NIs by the clinicians. e patients were assessed for SSIs, UTIs, respiratory infections, and BSIs as per the Center for Disease Control and Prevention (CDC) criteria [21]. Data were collected from selected patients using a questionnaire by five trained nurses. A questionnaire was developed from different kinds of literature [4,5,20]. Clinical specimens such as a clean-catch midstream urine, blood, wound swab, throat swab, nasal swab, and other body fluids were collected by trained laboratory technologists using standard procedures as described by Horan et al. [22] from the respective site of infections. e specimens were labeled with the patient's identification, packed and transported to the Bacteriology Laboratory of the Department of Medical Laboratory Sciences, College of Health and Medical Sciences, Haramaya University, within 30 minutes of the collection in a cold box for analysis.

Phenotypic Characterization of the Isolates.
In vitro phenotypic characterization of bacteria was carried out using standard culture and biochemical tests as described by Cheesbrough [23]. In brief, each specimen was streaked onto differential and selective culture media (Oxoid, Ltd., UK), including MacConkey agar, cystine lactose electrolytedeficient agar, mannitol salt agar, Pseudosel agar (cetrimide agar), and 5% sheep blood agar, for investigation of bacteria. e bacterial isolates were characterized based on colonial morphology, pigmentation of the colony, and cell morphology. Further characterization of the isolates to the species level was performed using biochemical tests, including coagulase, catalase, oxidase, urease, indole, citrate, glucose and lactose fermentation, triple sugar iron agar, bile solubility, and motility tests. Reading of the culture was done by two senior medical microbiologists after overnight incubation at 37°C. UTI was defined when patients had a fever (>38°C), urgency, frequency, dysuria, or suprapubic tenderness and a positive urine culture, i.e., ≥10 5 colonies forming unit per ml of urine with no more than two species of bacteria [22].

Antimicrobial Susceptibility Testing. A modified Kirby
Bauer disk diffusion method was used to test each isolate for in vitro antimicrobial susceptibility based on the Clinical and Laboratory Standards Institute criteria [24]. In brief, standard inoculum adjusted to 0.5 McFarland standard turbidity was uniformly distributed over the surface of Mueller Hinton agar (Oxoid, Ltd., UK). Antimicrobial disks, including (Oxoid, Ltd., UK) cephalexin (30 μg), ceftazidime (30 μg), ceftriaxone (30 μg), chloramphenicol (30 μg), erythromycin (15 μg), gentamycin (10 μg), penicillin (10 μg), tetracycline (30 μg), and cefoxitin (10 μg), were applied on Mueller Hinton agar plates using automatic disk dispenser. Following overnight incubation at 37°C, the zone of inhibition was measured and interpreted as sensitive, intermediate sensitive or resistant per the standard criteria [24]. Methicillin resistance was confirmed for all cefoxitinresistant S. aureus using a PBP2a-agglutination test as described by Atay and Gulay [25]. An isolate was considered multidrug resistant (MDR) if it is resistant to at least one agent in three or more antimicrobials of structurally different categories [26].

Data Quality Control.
e questionnaire was initially prepared in English and translated into local languages (Amharic and Afan Oromo) by a language expert and back to English by another language expert to check the consistency. e questionnaire was pretested on 5% NI-suspected patients in the Dil Chora Referral Hospital, Dire Dawa, Ethiopia, before data collection to check its simplicity. Data collectors were recruited from other hospitals in the region and trained for 4 days on the data collection procedure and interview techniques to minimize bias. Universal precautions were strictly followed during specimen collection, culture preparation, inoculation, and isolation. e sterility of the newly opened medium was checked before use.  e prevalence of NIs was presented in percentage along with the 95% confidence interval (CI). Two or more bacteria isolated from one patient were categorized as one bacterium for summarizing the prevalence but were analyzed separately for the antimicrobial susceptibility profile. Each bacterium was tested in triplicate for a single antimicrobial, and the mean value was taken to determine its antimicrobial susceptibility pattern. Intermediate results were included in the resistant category for analysis.

Ethical Consideration.
e study was reviewed and ethically approved by the Institutional Health Research Ethics Review Committee of the College of Health and Medical Sciences, Haramaya University. Informed, voluntary, written, and signed consent/assent was obtained from each parent/caretaker or participant. Positive finding was communicated to attending clinician for appropriate treatment.

Patient Characteristics and Prevalence of NIs.
A total of 394 clinically suspected patients for NIs were included in this study. e majority of study participants were females, 223 (56.6%), with a male-to-female ratio of 0

Antimicrobial Susceptibility Pattern.
Of the total Grampositive bacteria, 73.3% were resistant to tetracycline, 66.7% to erythromycin, and 53.3% to penicillin, while all other isolates showed sensitivity to antimicrobials in the testing panel. Of the Gram-positive bacteria, S. aureus showed resistance to a number of antimicrobials (80% resistance to each of chloramphenicol and erythromycin, 70% to cephalexin and tetracycline). S. pneumoniae isolates were 75% resistant to each of tetracycline and erythromycin, and 62.5% to penicillin. CoNS were 60% resistance to each of erythromycin and tetracycline. Streptococcus spp. showed 100% resistance to each of cephalexin, erythromycin, and tetracycline. MRSA accounted for 88.9% of a total number of S. aureus isolates ( Table 3).

Discussion
Nosocomial infections are one of the major public health problems around the world that vary from one country to the other. e prevalence of culture-confirmed NIs in this study was 6.9% (95% CI: 4.3-7.9). is was lower compared with a previous study conducted in a Tertiary Care Hospital, Ethiopia (35.8%) [20], and Rabat, Morocco (10.3%) [1], but higher compared to the study reported in Lambarene, Gabon (0.3%) [27], and Mazandaran, India (1.03%) [5]. e higher prevalence of NIs in the present study might be due to the inclusion of all age groups and a large number of different kinds of specimens from different wards; the other studies only looked at adult patients and focused on limited types of specimens. e prevalence of SSIs (31.5%) in this study was higher compared with a study reported from Bahirdar, Ethiopia (10.9%) [19], and Zahedan, Iran (8.6%) [14]. e higher prevalence of SSIs in this study could be explained by the lack of appropriate institutional strategies for the prevention and control of bacterial infections. It could also be due to the fact that this hospital services as a referral hospital that receives patients with severe injury (Mencha injury) from all over the eastern part of our country, which facilitates the circulation of pathogenic bacterial strains that cause SSIs. e prevalence of BSIs (25.9%) in this study was relatively comparable with the findings of a study conducted in Liaquat University Hospital, Hyderabad, Pakistan (22.7%) [28], and a Tertiary Care Hospital, Wollo, Ethiopia (20.8%) [20]. e finding was much higher compared with a study conducted in the Felege Hiwot Hospital, Bahirdar, Ethiopia   (2.4%) [19]. e higher rate of BSIs in this study might be due to the inclusion of patients from varying admission wards that increase the likelihood of getting an exogenous and endogenous infection of the bloodstream, while the previous study was limited to the intensive care unit and postoperative infections. e most frequent bacteria causing NIs were S. aureus (18.5%) and E. coli (16.7%). e preponderance of S. aureus in this study was in agreement with a study conducted in Bahidar, Ethiopia [19] and other countries [1,27]. is might be due to its association with the endogenous source as the organism is a member of the skin and nasal flora of the patients [23], cross-contamination from the hospital environment, surgical instruments, or hands of the health professionals [12,29]. e second most occurrences of E. coli in this study might also be due to the profound influence of endogenous contamination from the bowel, hollow muscular organ of patients, and transfer from an object to the human in a hospital setting resisting common antiseptics [23,30].
While most resistant Gram-negative and positive bacteria in this study were consistent with other findings [4,5,20], the resistance profile of certain strains to commonly used antimicrobials for the treatment of bacteria causing NIs was alarming. e higher prevalence of resistance observed in S. aureus to a number of antimicrobials (80% to each of chloramphenicol and erythromycin and 70% to each of cephalexin and tetracycline) is of a particular concern. A similar finding was reported in a tertiary hospital, Ethiopia [20], and other countries over the two decades [1,31]. Moreover, the resistance of S. aureus to methicillin (88.9%) in our study was higher compared with other studies conducted elsewhere: Tertiary Care Hospital, Pakistan (64.1%) [32], Gauteng Academic Hospitals, South Africa (36%) [33], and Canadian Hospital, Canada (26%) [13]. e high rate of MRSA observed in this study might be attributed to the lack of molecular assay targeting the mecA gene [34]. We used PBP2a-agglutination test, which was the far inferior method as compared to molecular assay, to confirm MRSA. It could also be due to the incorporation of varying clinical specimens, high turnover of health professionals, low specialist staffing, high patient load, and poor prevention measures, which facilitate the circulation of resistant bacterial strains. e isolation of MDR bacteria (42.6%) in this study was lower compared to a study conducted in Zahedan, Iran (95%) [14], Ghana (70%) [35], and Kampala, Uganda (58%) [36]. Although the rate of MDR in this study was low, it has serious implications on modern medicine and patient care outcomes. e low rate of MDR in this study might be due to lack of investigation of anaerobic bacterial NIs.
is study has several limitations. First, it was a single hospital-based study. Second, patients were not followed after discharge due to difficulties in follow up and communication.
ird, anaerobic bacterial NIs were not investigated due to limited laboratory facilities. ese limitations are likely to 17

Conclusions
In conclusion, the prevalence of culture-confirmed bacterial NIs in this study was comparable with other similar study findings. e most common infections were surgical site and bloodstream. S. aureus, E. coli, and S. pneumoniae were the most frequent causes of NIs. Most of the isolates were resistant to commonly used antimicrobials in the testing panel. e treatment and management of bacterial NIs need to be supported by culture isolation and antimicrobial susceptibility testing. Ciprofloxacin, ceftriaxone, and gentamycin should be used for the treatment of NIs when empiric treatment is unavoidable. Future studies are recommended to measure the true prevalence of NIs and antimicrobial resistance by including district hospitals, discharged patients, and communities.

Data Availability
e SPSS data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.