Urinary Tract infection, Drug Resistance Prole and Fetal Outcomes among Pregnant women in Selected Health Facilites of Addis Ababa, Ethiopia: A cross-sectional study design

Background: Urinary tract infection (UTI) is a commonly encountered health problem among pregnant women. Untreated UTI may result in fetal complications like preterm birth, low birth weight, intra uterine growth retardation and intrauterine fetal death. One of the major problems with bacterial pathogens is their ability to develop drug resistance. Determining drug susceptibility pattern of bacteria from urinary tract helps to identify effective drugs and minimize further adverse perinatal outcomes. The purpose of this study was to describe the magnitude and drug resistance prole of pregnant women and related fetal outcomes. Methods: A cross-sectional study was conducted from March to June 2019. Socio-demographic data of the study participants was collected by administering structured questionnaire after obtaining full consent of the participants. Clean catch mid-stream urine was collected from the study participants and the samples were transported to the laboratory with screw-capped container. Blood and MacConkey agar were used to cultivate bacteria from urine sample. Bacterial colonies were isolated and identied using standard biochemical properties. Antibacterial susceptibility test was done on Muller-Hinton agar using antibiotic discs. At the time of delivery, infant’s birth weight and gestational age was recorded on the log book as part of the daily work. Fetal outcome data were recorded from the log book and entered to EPI data management software then transferred to SPSS for analysis. Results: Out of 424 urine samples processed, 63 (14.9%) yielded signicant bacteriuria. Fifty-nine out of 63 (93.7%) cases of UTIs were recorded among age group of 15-34 years. Escherichia coli, Klebsiella pneumoniae, and Staphylococcus aureus were the three dominant bacterial isolates. Most of the gram negative bacterial isolates were resistant to ampicillin and ceftazidime (43.2% each) while gram positive was resistant for tetracycline (36.8%). The result of this study showed that UTI had statistical association with the occurrence of Intrauterine Fetal Death (IUFD) and Prerupture of Membrane (PROM) (P<.05). Conclusion: Adverse fetal outcomes are signicantly linked with the presence of UTI during pregnancy. The condition can

the mother during pregnancy [19]. Iqbal et. al. also reported that, 11.3% of the deliveries from UTI positive pregnant mothers produced low birth weighted babies [5].
If they are left undetected and untreated, they may cause adverse maternal and fetal outcomes [20]. To this end determining uropathogen isolates and their drug susceptibility pattern among pregnant women is of a highest priority before deciding to manage UTI cases [21]. Two factors-Poor patient compliance and incomplete course of antibiotic therapy are responsible for the evolution of resistant bacterial strains to most of antibiotics [3,22].
The distribution of antimicrobial susceptibility data of UTI causing microorganisms changes from time to time and from place to place [23]. This urges for the isolation of bacterial strains and their drug resistance pro le for that speci c location before treatment is given based on sign and symptoms of patients.
There should be periodic monitoring of etiologic agents of UTI and their resistance pattern to avoid further complication of the case [24]. Therefore, screening for UTI in pregnant women is very important if the prevalence rates are ≥ 2% [25].
The studies regarding uropathogens and their antimicrobial susceptibility patterns in pregnant women and the related fetal complications in Ethiopia are limited [26].
Studies have shown the strong association that exist between maternal UTI during pregnancy and fetal complications including low birth weight, APGAR score, PROM, and IUFD [5,18,19].
Thus; the result of this study provide vital information on type of bacterial isolates and the drug resistance pattern in relation to fetal complications in pregnant women attending at study sites.

Methods
All pregnant women on their third trimester during the study period and attending antenatal clinic at the two health centers, and Tikur Anbessa Speclised Hospital that ful l the inclusion criteria were the study population.
Four hundred twenty-four (424) participants were involved in the study by using single proportion formula.
Every pregnant mothers attending the antenatal clinic who were willing and available at the time of study, ful lling the inclusion and exclusion criteria were enrolled by non-probable consecutive technique until the number reaches the required sample size.
A structured questionnaire was used to collect socio-demographic data and clinical history of the study participants. Age, marital status, educational status, UTI sign and symptoms and questions related to fetal outcomes were included on the questionnaire. Participant's clinical data and fetal outcome data were taken from the individual card and ANC log books.
Midstream urine sample were collected from the study participants from March to June 2019. Participants were informed on how to bring "mid-stream" urine of 15-20 ml volume through the following process.
Thorough hand washes with water and dry Wash the genitalia Void the rst 20-30 ml of urine in the toilet then collect the remaining urine in to a sterile clean container.
Then, the samples were transported to Ethiopian Public Health Institute (EPHI) microbiology laboratory within an hour after it was collected. All urine specimens were collected in a well labeled, sterile, dry, wide-necked, leak proof and screw capped container.
A calibrated sterile wire and plastic disposable loops were used to inoculate 1 µl of each urine specimen on MacConkey and blood agar plates.
Identi cation of the isolated bacterial pathogens were done on the basis of culture morphology and biochemical characteristics. Indole test, citrate, urea, coagulase, catalase, motility agar, lysine iron agar and novobiosin disk were used for identi cation purpose.
Antimicrobial Susceptibility testing were performed on isolates according to the criteria of Clinical and Laboratory Standards Institute (CLSI, 2019) using the Kirby-Bauer disc diffusion method on Muller-Hinton Agar. A homogeneous bacterial suspension was prepared by mixing a loop full of pure bacterial colony in 5 ml of 0.85% sterile saline solution. After adjusting the 0.5 McFarland standard of the suspension, it was evenly inoculated on Muller Hinton agar plate using sterile cotton swab and antibiotic disks were placed at 15 mm and 24 mm distance from the edge and from each other respectively. The plates were then incubated for 24 hours at 37 o c. The isolates were categorized as sensitive (S), resistant (R) or intermediate (I) based on the inhibition zone they produced around the antibiotics used following the described break point in CLSI.
Con rmation of AST result was done by Vitek 2 Compact system. AST-GN73 and AST-GP74 cards were used for the susceptibility testing of gram negative and gram positive bacteria respectively.
The drugs that were used in the susceptibility test of gram negative bacteria are Ampicillin (10 µg), Ampicillin/Sublactam (10 µg), Piperacillin/ Tazobactam During data collection and sample collection, data quality was ensured through the use of standardized questionnaire, proper orientation and intensive supervision by the principal investigator. In addition, the questionnaires are translated in to Amharic language for easy understanding. SOPs of the microbiology laboratory were followed.
Sterility of the media was checked by overnight incubation of the media at 37 0 c prior to inoculation. Quality control bacteria were used to check the growth support of the media. For blood agar plate (S. pyogen, E. faecalis, S pneumonia), for Mackonkey agar plate (E. coli, P. mirabilis, E. faecalis) and for Muller Hinton agar plate (S. aureus, E. coli, E. faecalis) were used. All the data were checked for their completeness and representativeness prior to entry.
Findings of the study was further explained in words and tables. Binary and multiple logistic regression analysis was used to see if there is an association with different variables. In all cases P-value less than 0.05 considered as statistically signi cant.

Socio demographic characteristics
There were a total of 424 participants and the age distribution of study subjects ranges from 18-40 years with the mean (standard deviation) of 27 ± 4.6 years and 393 (92.7%) of them belongs to the age group 15-34 years.
It was observed that around 1.4% of the study participants were divorced. Only 41 (9.7%) of the study participants had higher education and 293 (69.1%) were housewives. Based on their gravidity, more than half of the study participants (62.3%) were multigravida and 80 (18.9%) of the study participants had history of previous urinary tract infection. None of the study subjects were cigarette or alcohol addicted (Table 1).

Magnituide of UTI and distribution of bacterial isolates
Four hundred twenty-four (424) urine samples were collected from pregnant women attending antenatal clinic at the study sites, out of which 63 (14.9%) were positive for the presence of signi cant bacteriuria ( Figure 1).
From the total study subjects, 80.7 % of them had no symptoms of UTI out of which 11.6 % had signi cant bacteriuria in their urine.

1.4.Fetal outcome and its association with UTI
From the total of 424 pregnant women there were 57 (13 %) and 367 (87%) poor and good fetal outcomes respectively. Poor fetal outcome includes the occurrence of either one or more of the following: APGAR score in 5-min (below 7), low birth weight (below 2500 gm), Premature rupture of membrane, Preterm labor, or IUFD. The presence of bacteriuria, had a signi cant association with fetal outcomes with (p value =0.000). Maternal age, gravidity, educational status, occupation, marital status and history of UTI had no association with the UTI status of the mother (Figure 3).
PROM had a signi cant association with the occurrence of maternal UTI (P=0.000). However, no signi cant differences were noted between pregnant mothers with or without UTI in terms of IUFD, birth weight, 5-min Apgar scores and preterm labor, (20

Discussion
The study was done to determine the magnitude and drug resistance pro le of bacterial species isolated from pregnant women during the time of their third trimester and to check the association between UTI and adverse fetal outcomes. The overall prevalence of signi cant bacteriuria in this study was 14.9% (63/424), the result correlates with the nding of similar study conducted in Dire Dawa (14%), Harar (15.5%), Gondar (10.4%) but lower than the studies conducted in Nigeria (85%), Benin (74%) and Goba and Sinana Woredas, Bale Zone, South-East Ethiopia, (35.3%) [6,12,[27][28][29][30]. Sample size, trimester of pregnant women and demographic difference might be mentioned as the possible reason for the variation.
The prevalence of UTI among symptomatic pregnant women in this study was lower than the prevalence of asymptomatic at a rate of 3.3% and 11.6% respectively. The prevalence of symptomatic bacteriuria in the study site was in agreement with the result from similar studies done in Gondar (3.4%), lower than the result from Harar (21%), Dire Dawa (8%), Yemen (66.6%) but slightly higher than the result from Bahir Dar (1.9%) [6,8,12,17,27].
With respect to asymptomatic bacteriuria, the nding from Harar, Ethiopia, (13.0%), matches with this study but lower incidence rate was reported in Bahir Dar, Ethiopia, (7.6%), Dire Dawa, Ethiopia, (5.9%), and Gondar, Ethiopia, (7.0%) [6,12,17]. Higher prevalence of asymptomatic and symptomatic UTI reported in Yemen might be attributed to the less sample size, geographical variation or may also be due to di culties in interpreting the symptoms by the study participants or the data collectors.
Sixty eight percent (68.75%) of the total bacterial isolates belongs to gram negative similar to the study conducted in Gondar (67.5), Dire Dawa (73.1), Yemen (63.4%), Tanzaniya (61.9%), but lower than the report from Harar (90.3%), and India (100%) [6,8,12,27,31,32]. In addition to the sample size variation, the environmental conditions, population difference and socioeconomic status of the participants can be mentioned as possible reason for the result variation.
With regard to bacterial isolates, E. coli was the most dominant bacterial isolate (44.4% and 63.6%) from the total and from gram negative isolates respectively. Similarly, high proportions of E. coli from UTI positive pregnant women were reported in Nigeria (66.6%), India (43.9%), Dire Dawa (34.6%), Gondar (47.5%), and Tanzania (33.3%) [2,6,12,31,33]. In most of these cases, either Kleibssiela spp or coagulase negative Staphylococcus are the second predominant organisms next to E. coli. In contrast to the above ndings, Citrobacter freundii were reported by Tula and Lyoha in Nigeria as the most dominant bacterial isolate from UTI positive samples [34].
In this study, K Pneumoniea and S. aureus were equally ranked as the second predominant pathogens while S. epidermidis and E. faecalis ranked third. These ndings indicate that even though there are a number of UTI causing bacterial species, majority of UTI cases are limited to few groups of bacteria.
Amikacin, cipro oxacin, gentamicin, cefoxitin, tobramycin and levo oxacin were the most effective drugs over gram negative bacteria by producing more than 90% susceptibility.
In the study conducted by Alemu et al., ceftriaxone, cipro oxacin, nor oxacin, gentamicin, amoxicillin-clavulanic acid, co-trimoxazole and tetracycline produced the highest level of sensitivity over gram negative bacteria [6]. Disk diffusion and MIC breakpoints for nor oxacin are deleted from CLSI M100, 29th ed. [35]. Similarly, ceftriaxone, gentamicin and cipro oxacin were the most effective drugs against gram negative bacteria in the study done by Derese [12].
In Libya, nitrofurantoin, amikacin, o oxacin, cipro oxacin and cefotaxime were reported as drugs that produced high susceptibility to most of the isolates [15].
From gram negative bacteria, E. coli showed sensitivity to all the antibiotics tested. On contrary, a study done in Mekelle, Northern Ethiopia, reported a resistance of E. coli to amoxicillin-clavulnic acid and trimethoprim / sulfamethoxazole and ampicillin [36].
Similarly S. aureus also showed resistance to tetracycline but was sensitive to most of the antibiotics tested, this result differs from the result reported in Mekelle [36].
When we compare the rate of isolates of Gram negative and Gram positive bacteria. Gram negative bacteria were the dominant causative agent of UTI which is in line with others report [12,15,36].
In this study, the association of low birth weight and maternal UTI were not signi cant (P = 0.521). About 16 (80%) and 4 (20%) underweight deliveries were recorded from 361 UTI negative and 63 UTI positive pregnant mothers respectively. This proportion of low birth weight were in line with the study conducted by Iqbal et al., However; in a study conducted by Amiri and his colleague in Ghana, birth weight were signi cantly associated with maternal UTI status [18,19,37].
The difference in sample size, difference in settings and sampling method can explain the difference in the association of the two variables in different geography.
Like birth weight, APGAR score (in the 1st and 5th minute), IUFD and gestational age were not statistically associated with maternal UTI; however, in a study conducted by Siakwa in Ghana, occurrence of low APGAR score (less than 7) and preterm labor respectively were 3.8 and 3.4 times than their occurrence in UTI negative mothers. A recent study done by Farah Iqbal also reported similar association of preterm delivery and low birth weight with UTI. A retrospective and observational study in Brazil further points out the signi cant association that exist between preterm birth and UTI but with respect to APGAR score, the result had similar nding with this study. The difference in signi cance level and odds of occurrence may be attributed to the socio-demographic difference of study subjects. [8,19,38].  [19,38].
According to different studies, UTI induces macrophages to release metalloproteinase enzymes that can degrade amniotic membranes and ultimately leads to rupture of amniotic membranes. [19,38].
Our ndings have some limitations, like the sample size though enough to describe the magnitude of UTI, it looks small while we are considering the fetal outcome. Second, being a cross sectional study, the exact roles of UTI in pregnant women and fetal outcome is confounded. If a follow up study were followed, we could understand the role of UTI in fetal outcomes during the course of pregnancy.

Conclusion
The nding from this study showed that the prevalence of UTI in the study area was high in general and higher proportion was taken by pregnant women without symptoms. Gram negative bacteria remained as the dominant causative agents of UTI. E. coli, S. aureus and K. Pneumoniae were the frequent isolates. The antibiotic susceptibility test showed variety of drug resistance pattern among different bacterial isolates. E. coli were susceptible for most of the antibiotics tested. Majority of gram positive isolates were resistant for Tetracycline and the two antibiotics: Tetracycline and Trimethoprim/Sulfamethoxazole were not effective against the most prevalent gram positive bacteria, S. aureus. UTI was found signi cantly associated with fetal outcome after keeping all other variables constant. Early screening of UTI in pregnant women is useful to minimize complication related to fetal outcomes. The exact roles of UTI in fetal outcomes during pregnancy needs further research works with better study design and sample size.

Informed consent was obtained from all participants
Proper explanation about the study was given to the study participants and the study was conducted after obtaining full verbal and written consent from pregnant women attending at antenatal clinics at the study sites.
The purpose of the study was explained to the study participants and also they were informed that the procedures used in the study do not cause any harm to them or their child. They were told that they have a full right to participate or not, to withdraw the consent and stop participation at any time without any form of prejudice.
Study participants were also informed that con dentiality of the information and privacy of the respondents will be assured by using codes instead of their names at each step of the study process. In addition, they were informed about the bene t of being part of the study. Positive UTI results were given to clinician and the study participants got appropriate treatment. All methods were performed in accordance with the relevant guidelines and regulations.

Figure 3
Fetal outcome of pregnant women, 2019.