Neonatal septicemia at intensive care unit, Ayder Comprehensive Specialized Hospital, Tigray, North Ethiopia: Bacteriological profile, drug susceptibility pattern, and associated factors

Background Neonatal septicemia is a life threatening medical emergency that requires timely detection of pathogens with urgent rational antibiotics therapy. Methods A cross-sectional study was conducted between March 2017 to September 2018 among 317 septicemia suspected neonates at neonatal intensive care unit, Ayder Comprehensive Specialized Hospital, Mekelle, Tigray, North Ethiopia. A 3 mL of blood was collected from each participant. Identification of bacterial species was done using the standard microbiological techniques. Antibiotic sensitivity test was done using disk diffusion method. Data were entered and analyzed using computer software SPSS version 22. Bivariate and multivariate regression analysis was applied to determine the association between variables. Results Of the 317 (190 male and 127 female) neonates, 116 (36.6%) were found to be with culture proven septicemia. Klebsiella species were the predominant etiologic agents. Length of hospital stay (AOR (adjusted odds ratio) = 3.65 (2.17–6.13), p < 0.001) and low birth weight (AOR = 1.64 (1.13–2.78), p = 0.04) were the factors associated with neonatalsepticemia. Most isolates showeda frightening drug resistance rate to the commonly used antimicrobial drugs. K. pneumoniae, E. coli, Enterobacter and Citrobacter species were 57% to100% resistant to ceftazidime, ceftriaxone, gentamycin, amoxacillin-clavulunic acid and ampicillin. All, 9 (100%) isolates of S. aureus were resistant to oxacilline, ampicillin,erythromycin and gentamycin. Furthermore, 55.6% S. aureus isolates were Methicillin Resistant Staphylococcus aureus. Conclusion Neonaltal septicemia is found to be significantly high in the present study. As most of the isolates are potentially related to hospital acquired infections, prevention and control policy should have to be more strengthening in the neonatal intensive care unit.

comprises about 6 million people) and other neighborhood regions such as Afar and Amhara. It also provides local emergency service.

Study design and study period
A cross-sectional study was conducted from March 2017-September 2018.

Sample size and sampling technique
Sample size was calculated based on a single population proportion formula as follows: Where; p = prevalence of bacterial isolates(32.1%) which was taken fromprevious report in Gondar, Ethiopia [12], d = degree of accuracy desired (0.05), Z 2 1α/2 = the standard normal deviation (1.96).
Finally, a total of 317 (94.6% response rate) participants were recruited using aconsecutive convenient sampling technique.

Requirement criteria
Inclusion criteria. After suspicion of neonatal sepsis is made by the attending pediatrician, babies with a clinical diagnosis of neonatal sepsis at the time of admission or during their hospital stay were enrolled in the study.
Exclusion criteria. Babies who had received antibiotics before admission or who had criticalcongenital anomalies were excluded from the study.

Data collection
A structured data collection format derived from the guidelines laid down by WHO young Infant Study Group [19] was used to obtain socio-demographic data and other relevant information such as Intrapartum fever, premature rupture of membranes (PROM), weight of the baby, gestational age, mode of delivery, presence or absence of lethargy, fever, failure to suck, moro reflex, tachypnea, apnea, tachycardia, respiratory distress, seizures, asphyxia and jaundice.

Laboratory procedures
About 3 ml of whole blood was collectedaseptically and inoculated into Brain Heart Infusion broth (BHI) in a ratio of blood: BHI of 1:10 and transported to the Microbiology laboratory for further processing. After 24 hrs incubation in Brain Heart Infusion broth, sub-culture was done onto blood agar, chocolate agar and MacConkey agar. Broth cultures with negative results in the first 24 hrs incubation were further re-incubated and then sub cultured after 48 hours, 96 hrs, with last sub-culture on day 7. Identification of bacterial colony was made bycolony morphology, gram staining reactions and biochemical reactions such as catalase, coagulase, hemolytic activity, triple sugar iron, indole, motility, citrate, urease and hydrogen sulphide production tests.

Data quality control
Laboratory analyses were carried out using standard operating procedures (SOPs). Prior to the actual work, reagents were checked for proper functioning and expiry dates. All culture media were prepared following the manufacturer's instructions. Each batch of the prepared media was checked for sterility by incubating a sample medium (5%) at 37 � C for 24hrs. Known bacterial species were inoculated and incubated at 37 � C for 24hr for the performance check. Escherichia coli ATTC 25922, Staphylococcus aureus ATTC 25923 and Pseudomonas aeruginosa ATTC 27853were used as control strains.

Data analysis
Data were analyzed using computer software (SPSS version 22). Frequency and percentage were employed to summarize the results and presented in tables. In univariate analysis, all variables witha p-value of < 0.05 were subjected to multivariate analysis. In multivariate regression analysis, to determine the association between variables, a p-value < 0.05 with a corresponding 95% confidence interval was considered as statistically significant.

Ethical considerations
Ethical approval was obtained from the Ethical Review Committee of College of Health Sciences, Mekelle University (ERC 0885/2016). Written consent was obtained from each guardian and official permission letter was written from the College of Health Sciences, Mekelle University to Ayder comprehensive specialized hospital neonatal intensive care unit (NICU). Resultswere communicated to respective physicians for beneficiary measures.

Study participants
In the present study, a total of 317 (190 male and 127 female) neonates were participated. Age of the participants ranged 1 to 28 days and mean was 6.94 (±6.42 standard deviation (SD)). Majorities (62.1%) of the neonates were below or equal to 7 days. Weight of the neonates was between 0.95-4.3 kilograms and the mean was 2.35 (±0.73standard deviation). Majority (53.3%) of the neonates were born below 2.5 kilograms. The mean gestational period was 36.42 (±3.38) and 40.4% (128/317) were born before 37 weeks of gestation. Around 4.7% (15/ 317) of the neonates were born at home. Of the total, 26.3% (85/315) of the neonates were born bycesarean section. Thirty four (10.7%) neonates died of their clinical illness (Table 1).

Neonatal sepsis
Of the total, 116 (36.6%) of the study participants were found to be with culture proven neonatal sepsis. There was no statistically significant difference in prevalence of neonatal sepsis among age groups (p > 0.05). The prevalence of neonatal septicemia was higher among females, and those who born at gestational period <37 weeks, at home and by spontaneous vaginal delivery but it was not statistically significant compared to their counter parts (AOR > 0.05), where as neonatal septicemia was higher among those whose weight was < 2.5 kgs at birth compared to those who born with > = 2.5 kgs and this was statistically significant difference (AOR = 1.64 (1.13-2.78); p = 0.04) ( Table 2).

Associated factors
Among the assessed risk factors, hospital stayshowed statistically significant association with culture proven neonatal septicemia: neonates who stayed for greater than or equal to three days in the hospital were about 3.7 times more at risk to develop sepsis compared to those who stayed for less than three days (AOR = 3.65 (2.17-6.13), p < 0.001). Whereas factors such as blood transfusion history, intrapartum fever, HIV status of the mother, prolonged duration of labour, PROM(Premature Rupture of Membranes), prolonged PROM and PPROM did not show significant association withculture positive results (p >0.05) ( Table 4).
With regard to the clinical presentations; jaundice, apnea and seizure showed statistically significant association with culture confirmed septicemia in binary logistic regression analysis (p<0.05). However, none of the indicated clinical signs and symptomsshowed statistically significant association with culture confirmed neonatal septicemia in multi-variant logistic regression analysis (p > 0.05). Culture confirmed septicemia did not show also significant association with the final prognosis of the neonate (Table 5).

Discussion
Neonatal sepsis remains a life threatening problem in intensive care unit especially in developing countries like Ethiopia. As the epidemiology and spectrum of the causative organismsof neonatal sepsis varies over time and from place to place, continuous assessment and
In the present study, majority of the isolates were Klebsiella species. Similarly, Klebsiella pneumoniae was the most frequently isolated organism from other previous studies: India [15,18,22,26] and Addis Ababa Ethiopia [14]. Though not a predominant, Klebsiella species were also among the most common pathogens reported in bloodstream infections from other previous findings in Ethiopia (13), and Egypt [16]. Klebsiellaspecies arenow becoming important public health concern especially in health care settingswith only few antibiotics to treat [27,28]. The high prevalence of Klebsiella species in the present study could be related to the ability of the organism to spread from patient to patient via contaminated hands of healthcare personnel or other persons, by contamination of the environment, or when the patients are on medical tools such as breathing machines and intravenous catheters. As Klebsiella species are ubiquitous in nature which can be found in the environment and on the mucosal surface of humans, the organism can be restrained in the hospital settings which enable to cause different infections among sick patients when there is lack of immunity, environmental and personal hygiene [28].
Presence of significant isolates of Citrobacter, Enterobacter, Acinetobacter and Serratia species was the other interesting finding of our study. As the infections caused due to these organisms are mostly nosocomial, the source of infection could be from the use of medical devices such as venous catheters, mechanical ventilators, or endogenously when the normal flora state alters due weakened immunity.
Of the assessed variables, length of hospital stay and the weight of the neonate showed significant association with presence of septicemia in the present study. Neonates who stayed greater as inpatient in the hospital were at high risk to develop bacterial septicemia and this may be related with the high probability for continuous exposure and acquisition of the causative bacteria from the hospital setting. Neonates who had low birth weight were at high risk to develop bacterial septicemia and the possible reason might be low birth weighted neonates are more likely to be with immature immunity and may have other co-morbidity that may easily expose them for secondary infections. The other possible reason could be low birth weighted neonates are mostly helped with different medical devicesin health care's and this might increase the risk of exposure to numerous infections. However, other socio-demographic and clinical related variables did not show significant association with blood culture proven bacterial septicemia. Nowadays drug resistance is becoming a major global challenge, and particularly in intensive care unit the rate is overwhelming because of the vulnerability of the patients due to reduced immunity, excessive use of broad spectrum antibiotics, invasive medical devices and procedures, and prolonged hospital stays. In the present study, most of the isolates were with frightening result to the commonly used antimicrobial drugs.
K. pneumoniae, the most frequently isolated bacteria, was 90% to100% resistant to the common antibiotics: ceftazidime, ceftriaxone, gentamycin, amoxacillin-clavulunic acid and ampicillin. Similarly, resistance rate of E. coli, and Enterobacter and Citrobacter species were 57% to100%, 75% to 100% and 80% to 100% respectively to the above mentioned antibiotics. The high resistance rate of these isolates could be attributed to the wide prescription habitof broad spectrum antibiotics in the intensive care unit and in the hospital as a general. In the present study, all isolates of S. aureus were resistant to oxacilline, ampicillin and erythromycin and gentamycin, and furthermore 55.6% S. aureus isolates were MRSA using cefoxitine as screening method.
WHO recommends ampicillin plusgentamicin for empiric treatment of neonates with suspected clinical sepsis [29]. Our study however, showed that these antibiotics are highly resisted  by most of the isolated bacteria and this could be due to wide and over prescription of these antibiotics. On the other hand, since the antibiotics like meropenem and amikacin showed better antibiotic activityagainst the isolated bacterial species, under regular monitoring,these are recommended to be used in the hospital for the treatment of neonatal septicemia. In general, the high prevalence and multi-drug resistant bacteria highlights the need for strengthening of infection prevention and control strategy as well as regular monitoring of antimicrobial resistance in the hospital and especially at the intensive care unit. The present studyalso highlights a specialdue attention to be given for Klebsiella species (particularly Klebsiella pneumoniae and Klebsiella oxytoca) as the frequency of these organisms is found significantly high.
Undergoing research with large sample size including in other hospitals will help see the bacterial distribution and drug resistanceprofile in the region, and then to develop local treatment guide for neonatal septicemia. Finally we recommended strengthening of hospital infection control and program to combat nosocomial infection such as the Klebsiella species dominantly isolated in this study.