Multi-drug resistance of blood stream, urinary tract and surgical site nosocomial infections of Acinetobacter baumannii and Pseudomonas aeruginosa among patients hospitalized at Felegehiwot referral hospital, Northwest Ethiopia: a cross-sectional study

Background Multi-drug resistant (MDR) Acinetobacter baumannii and Pseudomonas aeruginosa are major causes of nosocomial infections globally. They are the current World Health Organization critical priority pathogens for resistance, Antimicrobial resistance (AMR) surveillance and discovery of new antibiotics. However, there is paucity of data on nosocomial infections (NIs) caused by such superbugs in Ethiopia. Therefore, this study determined the magnitude and profile of nosocomial MDR A. baumannii and P. aeruginosa infections among patients hospitalized at Felegehiwot referral hospital, Northwest Ethiopia. Methods A cross-sectional study was conducted at Felegehiwot referral hospital from April 1 to July 31, 2018. A total of 238 patients with blood stream, urinary tract and surgical site NIs were enrolled conveniently. Either blood, urine and wound swab specimens were collected and processed using standard bacteriological procedures. A. baumannii and P. aeruginosa isolates were identified using standard bacteriological techniques and confirmed by automated Vitek2 Compact. Antimicrobial susceptibility testing on isolates was performed using the disk diffusion technique. The results were interpreted as per the standard zone sizes of Clinical and Laboratory Standards Institute.Chi-square test was done to determine associations among variables. P value < 0.05 was considered statistical significant. Results The median age of participants was 29 years. Overall,20(8.4%) of patients had nosocomial MDR A. baumannii and P. aeruginosa infections. The proportion of nosocomial MDR blood stream, urinary tract and surgical site infections were 13(8.9%), 5(8.3%) and 2 (6.3%), respectively. Patients with NI had lower mean age (24.9 years) (P = 0.035). All isolates of NIs were from patients with intravenous catheterization. The frequency of NI was 9(3.8%) for MDR A. baumannii and 11(4.6%) for MDR P.aeruginosa. A. baumannii and P. aeruginosa isolates were 100% MDR. All isolates of A. baumannii and P. aeruginosa were 100% resistant to ampicillin and piperacillin.A. baumannii isolates were 33.3 and 44.5% resistance against meropenem and ciprofloxacin, respectively while P.aeruginosa isolates revealed 36.4 and 45.5% resistance against ciprofloxacin and meropenem, respectively. Conclusions Health care associated infections of MDR A.baumannii and P. aeruginosa are critical problems in the study area. Therefore, urgent focused interventions required to contain the spreading of MDR NIs. Treatment of NIs for patients on health care should be guided by antimicrobial susceptibility testing.


Background
Multidrug-resistant (MDR) gram-negative bacterial infections are recognized as one of the major threats to global health. They are leading causes of nosocomial infections in the world [1][2][3]. NI is an infection occurring in a patient at the time of care in a hospital that was not manifest or incubating during admission but developed after 48 h of hospitalization [4].
The hospital environment contains a large number of immuno-compromised individual's and patients with diverse bacterial pathogens and normal microbiota [1,5]. NI accounted 7-10% prevalence in the world [5]. According to 2014 World Health Organization (WHO) report, 15% of all the hospitalized patients suffered from NIs [6].
Surgical site, blood stream and urinary tract infections are the most frequently reported types of NIs [1,7]. Urinary tract infection (UTI) which involves the bladder, ureter and kidney is the most common type of NIs accounting, 31% [8]. UTI caused by A. baumannii and P. aeruginosa are usually hospital acquired and related to urinary tract catheterization or surgery [9,10]. Surgical site infection (SSI) can involve the skin, subcutaneous tissue of the incisions, deep soft tissues of the incisions and any part of organ and spaces and accounted for 17% of NIs [8,10]. Blood stream infection is also one of the common NIs next to UTI and SSIs [10]. It can be acquired from catheter, secondary to another infection site, invasive diagnostic procedures or foreign body [10].
Nosocomial infections with MDR organisms are major global health issues [1][2][3]. They are very difficult for treatment and main causes of poor clinical outcome, morbidity, mortality, prolonged hospitalization and high health care costs [11,12]. The situation is true and urgent in Ethiopia.
The non-fermentative gram negative bacilli A. baumannii and P. aeruginosa have emerged as serious particular concern [13,14]. They are among the most common and serious MDR pathogens documented along with Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae and Enterobacter spp., collectively acronyms and considered as ESKAPE pathogens and superbugs, respectively [14,15].
Multi-drug resistant A. baumannii and P. aeruginosa survive in the hospital setting, transmitted easily between patients through the hands of health care workers [2,14]. Earlier findings elsewhere in the world reported that A. baumannii and P. aeruginosa commonly possess inherent resistance to antimicrobial agents through reduced permeability of the outer membrane, efflux pump systems, enzymatic inactivation and biofilm formation [15,16]. Thus, they are often resistant to almost all βlactams, aminoglycosides and quinolones [15,17].
Nosocomial isolates of MDR A. baumannii and P. aeruginosa complicated the treatment of infections and had adverse effect on clinical outcomes and increases patient treatment costs [18]. Factors such as antimicrobial drug overuse, prescription of drugs without susceptibility testing, self-medication and long duration of hospitalization are reported to the occurrence of MDR [11,19]. However, there is a scarcity of data on the burden of nosocomial MDR A. baumannii and P. aeruginosa infections in Ethiopia in general and study area in particular. Unavailability of local antibiogram data linked with self-drug prescription by patients and poor awareness on antimicrobial resistance are also big issues. Thus, the study aimed at determining the proportion of nosocomial MDR A. baumannii and P. aeruginosa infection among patients clinically diagnosed for NIs at Felegehiwot referral Hospital (FHRH), Ethiopia.

Study design, period and setting
A cross-sectional study was conducted from April 1 to July 31, 2018 at FHRH, Bahir Dar, Northwest Ethiopia. Bahir Dar is a capital city of the Amhara National Regional State, located approximately 565 km northwest of Addis Ababa. FHRH is a tertiary hospital that provides services for 5 to 7 million people. It has 684 health care and 166 administrative workers. The hospital has 466 beds for inpatient services. FHRH consists of 13 different wards including Surgical, Medical, Gynaecology and Intensive Care Units (ICUs). It has adult and neonatal ICUs with 8 beds. Bacteriological procedures were conducted at FHRH microbiology laboratory and further confirmatory identification of isolates and screening of MDR made using VITEK2 compact at Amhara Public Health Institute (APHI), Ethiopia. All patients admitted at FHRH and clinically presumptive for either of nosocomial urinary tract, post-operative surgical site and blood stream infection during the study period were the study population.

Sample size and sampling
The sample size for NIs was determined using Epi info version 3.5.1 (public domain software, www.cdc.gov) by considering 95% confidence level and marginal error (5%). A proportion of 0.83 MDR NIs taken from previous study in other parts of South East Ethiopia [16]. Thus, the total sample size was 238. All patients who were clinically presumptive for nosocomial blood stream, urinary tract and post-operative surgical site nosocomial infections were included conveniently until the required sample size was achieved.

Variables
Multi-drug resistant blood stream, urinary tract and surgical site nosocomial infections A. baumannii and P. aeruginosa were the dependent variables while demographic variables (age, sex, residence, educational status, occupation) and clinical data on co-morbidity, urinary catheterization, intravenous catheterization, duration of catheterization, previous history of antibiotics, previous history of surgery, duration of hospitalization, wards of patients hospital and duration of operation were the independent variables.

Inclusion and exclusion criteria
All patients who had either clean and cleancontaminated operations or other medical reasons admitted in Medical, Surgical and ICU wards of FHRH and developed clinical evidences of blood stream, urinary tract and surgical site nosocomial infections after 48 h of admission were included during the study period. However, patients who had re-operation, contaminated and dirty operations, ventilator-associated pneumonia, skin infections and developed any type of clinical evidences of infection within 48 h of admission were excluded from the study.

Data collection
Information on demographic variables was collected from each participant by face-to-face interview using a structured questionnaire. Clinical data related to comorbidities, hospitalization, surgery and use of antibiotics were collected by reviewing patient's medical record and in consultation with the respective physician. With the study team, patients admitted in the Medical, Surgical and ICU wards were carefully diagnosed for NIs by Internists and Surgeon. Clinical specimens (blood, urine and wound swab) were collected by the study team as soon as NI was reported following the bacteriological standard procedures.
The criteria of The European Centre for Disease Prevention and Control were used to define NIs [10]. Patients who had either purulent drainage, pain, localized swelling, redness, or heat in the skin, subcutaneous tissue, deep soft tissue, organ or spaces and one positive culture for A. baumannii and P. aeruginosa after 48 h of operation were considered as nosocomial surgical site infection. Patients who had either fever (> 38°C), urgency, frequency, dysuria, or suprapubic tenderness with no other recognized cause but has ≥10 2 CFU/milliliters (ml) and ≥ 105 CFU/ml of urine culture for catheterized and non-catheterized patients, respectively after 48 h of admission were considered as nosocomial urinary tract infection. On the other hand, patients who had either fever (> 38°C), chills, or hypotension and one positive blood culture for A. baumannii and P. aeruginosa after 48 h of admission were considered as nosocomial blood stream infection.

Wound swab collection and processing
Two wound samples from each participant were collected aseptically by sterile cotton swabs dipped in normal saline using Levine method [20]. The wound swabs were inoculated on MacConkey agar (MAC) and Blood agar (BA) (Oxoid, UK) at a time. Both plates were incubated at 37°C and examined for visible bacterial growth after 48 h of incubation.

Blood sample collection and processing
As per the standard protocol, 1, 2 and 10 ml of venous blood were collected from neonates, children and adults presumptive for blood stream infection, respectively [21]. Following cleaning of the site of blood collection with 70% alcohol and 2% tincture iodine, two blood samples were collected from two different sites of peripheral vein of each febrile patient using two bottles of blood sample within 30 min difference. The collected blood samples were inoculated directly to 5 -10 ml Tryptic Soya broth blood culture medium bottle (Oxoid, UK) and transported to FHRH microbiology laboratory. All blood culture broths were incubated aerobically at 37°C and regular subcultures were done on BA and MAC (Oxoid, UK) after 1, 2 and then 3 days later daily up to 7 days of incubation. All of the inoculated agar plates were incubated at 37°c. Finally, plates were examined for bacterial growth after 24 h.

Urine sample collection and processing
Clean-catch mid-stream urine was collected from patients presumptive for UTI. From catheterized patients, 5 ml of catheterized urine transferred to a sterile container after cleansing the out let of catheter. For noncatheterized patients, the same amount of urine sample was collected by the patient using sterile container. Urine samples were inoculated on MAC and BA (Oxoid, UK). All agar plates were incubated aerobically at 37°C for 24 h and observed for bacterial growth. Blood agar colonies were counted using colony counter and checked for significant bacteriuria. Culture from catheterized and non-catheterized patients that grew ≥10 2 CFU/ml and 10 5 CFU/ ml, respectively was taken as a significant bacteriuria, respectively. For heterogeneous colonies, sub-culturing of individual distinct colonies was performed to ensure pure cultures.

Identification of bacterial isolates
All isolates of A. baumannii and P. aeruginosa were identified by manual standard microbiological methods [21]. All A.baumannii and P. aeruginosa suspected isolates were also further confirmed by an automated Vitek2 Compact (BioMérieux, France).

Quality control
The prepared questionnaire was checked for its completeness and validity prior to the data collection. All the standard operating procedures (SOPs) were strictly followed at all stages of microbiological analysis. American Type Culture Collection (ATCC) standard reference strains (P. aeruginosa ATCC27853 and E. coli ATCC 25922) were used for quality control of antimicrobial susceptibility testing. MDR A. baumannii and P. aeruginosa isolates were confirmed by using Vitek2 compact. A standardized bacteriological procedure was followed to maintain correct laboratory results. At regular intervals and whenever a new batch of strain or reagent is prepared, standard strains of P. aeruginosa ATCC27853 and E. coli ATCC 25922 were used as positive controls. The sterility of the media was checked by incubating the media overnight before its use. The data were checked for completeness and representativeness prior to entry.

Data analysis
Data were checked, entered and analyzed using Statistical Package for Social Science 23 (IBM Corp Released 2011.IBM SPSS statistics. Armonk, NY: IBM Corp). Descriptive statistics were used to describe relevant variables. Chi-square test, Fishers exact test and Independent samples T Test was obtained to determine association between dependent and independent variables. P value of < 0.05 was considered statistical significant.

Demographic characteristics
A total of 238 patients with clinical evidence of nosocomial infection (BSI, UTI and SSI) were enrolled in the study. Of them, 129 (54.2%) were males. The majority (21.4%) of participants were found in the age group of > 51 years with median age of 29 years. One hundred twenty six (52.9%) of the study participants were from urban settings. Data on occupation showed that majority (39.1%) of participants were government employee. Table 1 depicts the demographic characteristics of the study participants.
Participants with NIs had lower mean of age (24.9 years) than those without NIs (29.6 years) and the difference was statistical significant (P = 0.035). Highest (15.4%) NIs rate was found in age groups < 10 years. The proportion of NI was 13 (11.4%) in those participant with co-morbidity. All isolates of NIs were from patients with intravenous catheterization ( Table 3). Duration of operation was significantly higher in those patients with confirmed NIs (180 min) than their counterparts (155 min) (P = 0.04). Moreover, the duration of catheterization was higher in those with confirmed NIs (13.6 days) than their counter parts (11.3 days). However, the difference was not statistical significant (P = 0.25) ( Table 3).
Multi-drug resistance profiles of A. baumannii and P. aeruginosa isolates A. baumannii and P. aeruginosa isolates were resistant to three to six antibiotics from different classes. All isolates of A. baumannii (9) and P. aeruginosa (11) were MDR (100%). Among this, 3 (33.3%) isolates of A. baumannii and 4 (36.4%) isolates of P. aeruginosa showed resistance to antibiotics from six different classes, respectively. Three (33.3%) isolates of A. baumannii and 1 (9.1%) isolates of P. aeruginosa were resistant against antibiotics from 4 different classes, respectively. Moreover, 3 (33.3%) of isolates of A. baumannii and 6 (54.5%) of P. aeruginosa isolates were resistant against antibiotics from three different classes (Table 4).

Discussion
Antibiotic resistant nosocomial infections are becoming serious health care problem in ICU and other areas of hospital care, increasing morbidity, mortality, length of stay, and health care costs [2,24]. The epidemiological and antimicrobial resistance profiles of NIs showed variations among hospitals around the globe. Many of the infections are caused by bacteria that are resistant to multiple antibiotics [24,25]. This study showed the proportion of NIs due to two MDR non-fermentative gram negative bacilli among patients hospitalized in different wards of a referral hospital.
In the present study, 8.4% of patients were infected with nosocomial MDR A. baumannii and P. aeruginosa. This indicated that MDR A.baumannii and P. aeruginosa   infections are the major health problem in the clinical area in Ethiopia. High patient load, overcrowding, poor infrastructure, poor infection control practices of the hospital and differences in trained medical staff for aseptic procedures might be the possible explanations. This finding was coherent with reports in Tikur Anbessa Hospital, Ethiopia (8.12%) [25], Uganda (7.39%) [26], Italy (9.3%) [27] and Gaza city (6.9%) [28]. However, it was higher compared to reports from Hiwot Fana Hospital, Ethiopia (0.5%) [8], Gabon (5.7%) [29], China (0.78%) [30] and Indonesia (3.5%) [31]. In contrast, the overall nosocomial MDR A. baumannii and P. aeruginosa infections in the present study was lower than studies done in Ghana (23.5%) [32]. This might be due to variation in sample size, clinical site of infection, age of patients, hospital setting, duration of hospitalization, patients exposure to high risk devices or surgical procedures, microbiological methods employed for screening of MDR resistant strains. The highest proportion of MDR non-fermentative gram negative bacilli NI infection among the lower age groups in the present study is consistent with earlier studies elsewhere in the world [29,33,34]. On the other hand, in the present study all NIs observed among patients with intravenous catheterization. The rate of nosocomial infections was also significantly higher among patients who had prolonged time of operation than their counter parts. This was consistent with previous study in Tikur Anbessa Hospital, Ethiopia (25). This might be due to the high rate of exposure of patients to the two MDR pathogens from the hospital environment, health   #T: Number of isolates tested, R%: Percentage of resistant isolates care professionals, multiple invasive device and crosscontamination among patient's procedures. In this study, the proportion of nosocomial MDR A. baumannii and P. aeruginosa surgical site infection (6.3%) was comparable with previous reports from other part of Ethiopia (6.6%) [25] and Ghana (8.5%) [32]. However, it was lower than studies from Southeast China (28.5%) [35]. This could be the difference in the type of surgery, handling of surgical equipments and age of study participants as the present study included patients of any age groups.
The proportion of nosocomial urinary tract infection linked with MDR A. baumannii and P. aeruginosa (8.3%) isolates in the present study was consistent with a study conducted in Kenya (9%) [36]. However, it was lower than studies in Gabon (52.8%) [29], Indonesia (16.5%) [31] and United States of America (USA) (16%) [37]. This could be due to differences among study participants, catheterization and hospitalization.
The prevailing proportion of nosocomial BSI in this study (8.9%) was comparable with studies done in USA (10%) [37]. However, it was higher than studies from Indonesia (3.5%) [31]. In contrast, lower findings were documented elsewhere in Africa (20-70.3%) [29,38] and Southeast China (46.1%) [35]. The observed difference might be due to non-sustainable infection control practices in hospitals, difference in use of invasive medical devices, procedures, hospital type and diverse nature of study participants.
In this study all isolates of MDR P. aeruginosa were resistant for ampicillin and amoxacillin-clavulanic acid. This was consistent with reports from Tikur Anbessa hospital, Ethiopia (25) and Southeast China [35] where 87.5 and 100% resistance rate against ampicillin and amoxacillin-clavulanic acid respectively were noticed. Moreover, all isolates of P. aeruginosa in the present study were resistant against piperacillin. This was significantly higher than studies from Italy (25%) [27], Southeast China (12%) [35], Turkey (28.7%) [40] and Taiwan (66.8%) [41]. This might be associated with differences in the number of MDR strains of P. aeruginosa and patient type. The frustrating level of resistance against piperacillin antibiotic is an alarm for treatment to be guided with antimicrobial susceptibility testing as it was not prescribed in the study area (FHRH).
In this study, high levels of resistance to cephalosporins (cefotaxime (63.6%) and ceftazidime (100%)) were obtained against P. aeruginosa isolates. This was coherent with studies in Uganda [23], India [39] and Taiwan [41] where 71-77% resistance against ceftazidime reported. Moreover, 70.8 and 92.8% level of resistance against cefotaxime were documented in Sodo, Ethiopia [18] and Southeast China [35], respectively. On the other hand, low level of resistance against ceftazidime reported in other parts of Ethiopia (12.5-29.1%) [18,25] and Italy (31%) [27]. The highest level of resistance against third generation cephalosporins might be linked with excessive, mis and inappropriate use of these antibiotics in the area that drives selective pressure and emergence of MDR.
The resistance rate of P. aeruginosa isolates against meropenem (45.5%) in the present study was coherent with other studies in Ethiopia (41.7%) [18] and Asia (36.6-54%) [33,39]. However, higher level of resistance against meropenem was documented in Taiwan (73.2%) [41] and Saudi (81.8%) [42]. The relatively lower proportion resistance against meropenem in the present study might be linked with the nonexistence of meropenem prescription practice in the study hospital. In contrast, lower level of resistance against meropenem was reported in Uganda (14%) [26] and Turkey (20.4%) [40]. This could be due to variation in the availability of meropenem in each localities, prescription difference, misuse and inappropriate use of antibiotics.
In this study, all isolates of MDR A. baumannii were resistance against ampicillin. This was parallel with studies conducted in Tikur Anbessa Hospital (88.2%), Ethiopia [25] and Southeast China (100%) [35]. Moreover, high level (88.9%) of MDR A. baumannii isolates resistance to amoxacillin clavulanic acid in the present study was comparable with earlier studies in Nigeria [43] and Southeast China [35], where all isolates of A. baumannii were resistant against amoxacillin-clavulanic acid.
All isolates of A. baumannii were resistant to piperacillin in the present study. This was comparable with studies conducted in Southeast Asia (83.7%) [35] and Italy (81) [27]. Similarly, A. baumannii isolates also revealed 77.8 and 88.9% resistance against ceftazidime and cefotaxime, respectively which was comparable with studies done in Tanzania [44], Southeast China [35] and Italy [27], where resistance against ceftazidime and cefotaxime reported in 71-97.1% and 54.4-100% of isolates, respectively. Furthermore, 33.3% of A. baumannii isolates from the present findings showed resistance against meropenem. The finding was relatively similar with studies conducted in Sodo, Ethiopia [18], Tanzania [44] and Nigeria [43] where, 30.2-40% resistance rate against meropenem reported. However, in Saudi [42], resistance against meropenem was reported in 90.5% of A. baumannii isolates, respectively.
The high MDR proportion observed in two nonfermenter gram negative bacilli in this study is probably related to the contaminations and cross transmission of these bacteria from hospital environment [18], hands of healthcare workers, frequent use of broad spectrum antibiotics, inherent resistance nature to many antimicrobial agents and the ability of pathogens persist in the environment, and on medical devices for a longer time [2]. Moreover, high prescription practice of common antibiotics including third generation cephalosporins and use of drugs outside the hospital might contribute for the high resistance rate of A. baumannii and P. aeruginosa to different classes of antibiotics.
This study was limited to participants admitted in hospital but nosocomial infection that arose after discharge was not detected. The single-centre nature of the study introduced local practice bias. The study also unable to demonstrate cross-transmission of pathogens and ignored the effect of duration of antibiotic therapy at individual patient level. Moreover, certain potential confounders could not be included in the analysis.

Conclusions
Alarming proportion of nosocomial MDR A. baumannii and P. aeruginosa infection obtained in the study area.
All A. baumannii and P. aeruginosa isolates were resistant for at least three antibiotics representative of different classes. Therefore, urgent intervention towards nosocomial infection prevention practices required and treatment of patients on care should be guided with antimicrobial susceptibility testing.