Bacteriological Prole, Antibiotic Susceptibility Pattern and Predictors of bacteremia among Children with Bloodstream Infection

Background: Bloodstream infections are major causes of morbidity and mortality among children in Sub-Saharan countries. This study aimed to determine the prevalence of bacteremia, etiological agents’ antibiotic susceptibility pattern and predictors of bacteremia among children with bloodstream Infection. Methodology: This hospital-based cross-section study involved children aged two months to thirteen years. All children meeting the inclusion criteria during the study period were enrolled. All consenting parents were interviewed via a questionnaire to collect data, followed by a thorough physical assessment and venipuncture was done to collect blood samples. Data were analysed using SPSS version 23. Results: Among 242 study participants, 154(63.6%) were male and blood cultures were positive in 37(15.3%). Gram-positive and gram-negative bacteria constituted 32(80%) and 8(20%), respectively. The frequent pathogen found was Staphylococcus aureus 25(62.5%), followed by Enterococcus spp. 4(10%), Escherichia coli 4(10%), Pseudomonas aeruginosa 3(7.5%), Streptococcus pyogenes 3(7.5%) and 1(2.5%) Klebsiella pneumonia. The majority of bacterial isolates showed high resistance to commonly used antibiotics in the study area. Predictors of bacteremia were severe malnutrition, hydrocephalus, hyperglycemia, lethargy and BSI with no foci of infection. Conclusion: Prevalence of bacteremia was 15.3%. Gram-positive bacteria were more prevalent than gram-negative bacteria. Staphylococcus aureus and Escherichia coli were isolates BSI. Effective antibiotics for both gram-negative & gram-positive are imipenem, meropenem and piperacillin-tazobactam followed by amikacin; vancomycin & clindamycin for gram-positive organisms. curb antimicrobial see in and other studies, continuous antimicrobial failing


Introduction
Bloodstream infection (BSI) is the presence of bacteremia associated with infection. [1] Bacteremia signi es the presence of bacteria in the bloodstream. The bacteria can originate from intravascular or extravascular sites, and the bacteremia can be transient, intermittent, or continuous. [2] BSI due to different bacteria are common causes of morbidity and mortality globally. They require immediate and adequate antimicrobial treatment. [3,4] BSI are potentially life-threatening, requiring rapid identi cation and antibiotic susceptibility testing (AST) of the causative pathogen. Morbidity due to BSI ranges from self-limiting infections to life-threatening sepsis that requires rapid and aggressive treatment. [5,6] Blood cultures remain the 'gold standard' technique for the detection of bacteremia and diagnosis of BSI. [7] There is wide variation in the prevalence of BSI depending on the settings. Analysis of 3067 samples from children aged 0-15 who attended acute care in Switzerland, and whose samples were submitted to the Swiss antibiotic resistance centre (ANRESIS) lab network, yielded a prevalence of 59%. [3] In single centre studies, the prevalence of BSI is often lower. In a tertiary hospital Kathmandu Nepal, the prevalence was found to be 6.7% among 574 children aged 1-14 years clinically suspected of having BSI.[6] In Bangladesh, Tauhid et al. looked at 110 children under 5 in intensive care with suspicion of septicaemia and the prevalence of BSI to be 20.9%. [8] With similar criteria in Addis Ababa Ethiopia, they found a prevalence of 27.9% when looking at 201 children aged 0-12 years. [5] In Tanzania, a 2013 study in Mwanza reported a prevalence of 6.6% in 317 febrile children aged 1month -7years. [9] The problem is common in developed nations, but the burden is higher in Sub-Saharan countries where it carries a mortality of up to 53% in children. [10] Mortality due to BSI in children is as high as 22.4% in Dar es Salaam, Tanzania. [11] Inappropriate empirical treatment of BSI may increase mortality and encourage the emergence of drug-resistant strains. [12] The evolving etiology of BSI may affect the ability of clinicians to initiate timely, appropriate empirical antimicrobial therapy. Given the potential consequences of BSI and the changing AMR patterns of the causative agents, it is crucial to carry out investigations to con rm the distribution and burden of BSI. This will allow targeted prescribing of narrow-spectrum antibiotics to avoid driving further resistance towards the limited antibiotic options available. We aimed to describe the bacterial etiology, antibiotic susceptibility pattern and predictors of bacteremia among children with suspected BSI at a referral hospital in the hope that it will give light on to the current patterns in the region and help with better antimicrobial stewardship.

Study Design and Setting
This was a hospital-based cross-sectional study, conducted in the department of Paediatric and child health at Kilimanjaro Christian Medical Centre (KCMC) from January 2019 -May 2019. KCMC is the zonal referral hospital for North-Eastern Tanzania which has a population of 15 million people. The Paediatric department provides services for children aged 0-13 years.

Study Population and Sampling method
This study included all children aged 2 months to 13 years, presenting with two or more signs of systemic in ammatory response syndrome (SIRS). Signs of SIRS include either abnormal temperature (> 38.5°C or < 36°C) or abnormal leukocyte count; and either tachycardia or tachypnea for age.
All children admitted to the paediatric ward during the study period, who met the inclusion criteria were identi ed and enrolled in the study following parental consent. Using a previously reported prevalence of positive blood culture of 6.6% among paediatric patients from Bugando referral hospital [9]; the minimum sample size required for this study was estimated to be 162 children. We managed to enroll 242 children.

Data collection and laboratory procedure
Using a datasheet, we captured all the pertinent demographic information, clinical characteristics of the participants and the ndings of the physical assessment, which was conducted by the principal investigator. After that, we aseptically collected 5ml samples of blood for culture, full blood picture (FBP), as well as for RBG and HIV tests which are done routinely on all child admissions at KCMC. A laboratory data sheet recorded the results of laboratory ndings. We initiated empirical antibiotics following the collection of blood samples and adjusted the treatments after blood culture results.
For each of the participants, 3ml of blood was inoculated into a Paediatric BacT/ALERT® PF blood culture bottle (BioMerieux, Durham) and incubated in an automated BacT/Alert system (BioMerieux, Durham) at 35°C aerobically for seven days. If no growth was detected within seven days, the culture was recorded as negative. A positive blood culture identi ed by the system was further processed aseptically, by gram stain, with the gram-positive organisms subsequently being sub-cultured onto chocolate agar, blood agar and gram-negative onto MacConkey agar (Becton, Dickinson, USA). We then incubated the agar plates at 37°C for 24 hours before bacterial isolation. The growth obtained was identi ed by conventional biochemical tests for gram-positive and analytical pro le index (API) 20E test for gramnegative. The remaining 2 ml blood was used for FBP, RBG and HIV tests.
All isolates resistant to third-generation cephalosporins were tested for extended-spectrum betalactamase (ESBL) production using the disc approximation method. A diffusion disc of amoxycillin/clavulanate (20/10µg) was placed equidistance between a ceftazidime (30µg) disc and a cefotaxime (30 µg) disc on Muller-Hilton agar (Becton, Dickinson, USA). An enhanced zone of inhibition towards amoxycillin/clavulanate (20/10µg) disc was considered an indicator of ESBL production with a positive result de ned as a ≥ 5 mm increase in the zone diameter. [13] Methicillin resistance in Staphylococcus aureus (MRSA) was detected using oxacillin (1µg) discs by diffusion method on the bacterial lawn culture of Staphylococcus aureus on Muller-Hilton agar (Becton, Dickinson, USA). The zone of inhibition was determined after 24hour of incubation at 35°C. A zone size was interpreted according to CLSI (2017) criteria of ≥ 22 mm was considered sensitive, and ≤ 21mm was considered resistant. [13] For internal quality control of disc diffusion tests, ATCC control strains of E. coli (ATCC 25922), Staphylococcus aureus (ATCC 25923), Klebsiella pneumoniae (ATCC 13883) and Pseudomonas aeruginosa (ATCC 27853) strains were used for identi cation and assessing the adequacy of antibiotic discs for AST. Organisms which showed resistance to three or more different antimicrobial classes were considered multidrug-resistant (MDR) bacteria. [14,15]

Data analysis
To avoid the assessment of multiple outcomes for a single patient, only the rst admission per patient was included in this study. Data were entered, coded and cleaned before analysis. Data analysis was performed using SPSS version 23. Continuous variables were summarized using measures of central tendency (median) with respective measures of dispersion (Interquartile Range) while categorical variables were summarized using frequencies and percentages. Univariate analysis and multivariate logistic regression models were tted to determine the predictors of culture positivity. The ρ-values 0.05 were considered statistically signi cant.

Results
During the study period (January 2019 and May 2019), a total of 454 infants and children were admitted to the KCMC Paediatric ward. Of this, 242 (53.3%) met the inclusion criteria and nonrepetitive blood samples from clinically suspected cases of children with BSI were obtained within 24 hours of presentation to the pediatric ward, summarized in Fig. 1.

Predictors of bacteremia
In this study, severe malnutrition, hydrocephalus, hyperglycemia, lethargy and BSI with no foci of infection were found to be statistically signi cant predictors of bacteremia (Table 5 and Table 6).

Discussion
The overall prevalence of bacteremia in our study population was 15.3%. Gram-positive bacteria (80%) were more prevalent than gram-negative bacteria (20%). Escherichia coli (10%) and Staphylococcus aureus (62.5%) were the major gram-negative and gram-positive organisms causing BSI. The AST pattern of gram-positive and gram-negative bacteria showed low levels of resistance to amikacin (2.7%), piperacillin-tazobactam (3.3%) and vancomycin (3.4%). The strong predictors of bacteremia were severe malnutrition, hydrocephalus, hyperglycemia, lethargy and BSI with no foci of infection.
Our prevalence of bacteremia (15.3%) concurs with other studies performed in Tanzania (13.1%) and (13.4%). [11,16] Our ndings of BSI were lower when compared to other studies done in Switzerland (59%). [3] The lower rates of BSI may be ascribed to antibiotics being given prior to referral, either by clinicians in referring facilities or by self-prescription as antibiotics are available over the counter in Tanzania, and self-medication occurs often. Other studies have also found Staphylococcus aureus and Escherichia coli to be the prevalent organisms in BSI. [10,16,18] This suggests that infections by these agents constitute a substantial threat to child survival in this locale and other developing country settings. It also highlights the signi cant morbidity of community-acquired BSI in our region.
In the management of BSI in children, empirical antibiotic therapy should be unit-speci c and determined by the prevalent spectrum of etiological agents and their AST pattern. Among gram-positive organisms, Staphylococcus aureus was found to be resistant to penicillin 25/25(100%), followed by ampicillin 18/25(72.0%) and was sensitive to ceftazidime, amikacin, meropenem, imipenem, cefotaxime, oxacillin and piperacillin-tazobactam. This nding was consistent with other studies. [5,6,8,15] A similar pattern of antibiotic resistance, signifying that this is a problem far beyond our centre.
Among gram-negative organisms, Escherichia coli isolates were universally resistant to ampicillin and 75% resistance to amoxicillin/clavulanate, co-trimoxazole, and ceftriaxone. Nepal. [15] For MDR gram-negative bacilli, a combination of a third-generation cephalosporin (cefotaxime or ceftazidime) with amikacin may be appropriate but repetitive use of these antibiotics may increase the risk of ESBL infection. [19] This is of particular concern to clinicians since they have to treat children with invasive disease more cautiously to reduce morbidity.
Overall, resistance to antibiotics was as high as 89.2% in our study. Resistance was mostly seen to penicillin, ampicillin and co-trimoxazole in almost all organisms; similar to what has been reported in other studies. [16,17,20] In concurrence with studies elsewhere, we found gram-negative organisms to be more sensitive to cefotaxime, ceftazidime, piperacillin-tazobactam, amikacin, imipenem and meropenem. [15,17,19,21] We also similarly found gram-positive organisms to be sensitive to vancomycin, amikacin, piperacillin-tazobactam, meropenem and imipenem, in concurrence with these other studies. [15,17,19,21] The ndings perhaps could be due to the emergence of AMR with irrational use of empirical antibiotics contributing to its acceleration. Without clear efforts to curb the irrational use of antibiotics, we can anticipate resistance extending even to those antibiotics currently seen as sensitive.
This study showed that severe malnutrition, hydrocephalus, hyperglycemia, lethargy and BSI with no foci of infection were strong predictors of bacteremia, in concurrence with these other studies found severe malnutrition and lethargy predictors of bacteremia. [5,22] This highlights the multifaceted nature of bacteremia and the possibility that commodi es may play a signi cant role in its occurrence, presentation and prediction.

Strengths and limitations
This study of BSI offered a glimpse into resistance and sensitivity patterns to common antibiotics used in our setting, and the qualitative data plus the study successfully replicated other similar studies, thus answering a pertinent question.
The frequent use of antimicrobials prior to blood culture may have hindered the detection of organisms susceptible to commonly used antibiotics. Thus, the study results are likely to underestimate the prevalence of BSI.

Conclusions And Recommendations
The prevalence of bacteremia was 15.3%. Gram-positive bacteria were more prevalent (80%) than gramnegative bacteria (20% Availability of data and materials The data used and analysed during this study is not publicly available to guarantee participants privacy.
The data used to support the ndings of this study are available from the corresponding author upon request.

Competing interests
The authors declare no competing interests.
Authors' contribution SNC: conception and design of the study, statistical analysis, interpretation of data, manuscript writing, reviewing and overall accountability for all aspects of the work. MJM: statistical consultation and analysis, substantially reviewed the manuscript and accepting accountability for all aspects of the work.
LM, DM, RMM and RNP: critically reviewing the article; gave nal approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.