Variation in Progesterone Levels and Urinary Tract Infections in Pregnant Women Attending Moi Teaching and Referral Hospital, Eldoret, Kenya ()
1. Introduction
Urinary tract infection (UTI) defined as the presence of uropathogens greater than 105 CFU/mƖ in urine culture is a common health problem among pregnant women [1]. UTIs can present symptomatically [2] or asymptomatically [3], usually affecting 18% and 13% pregnant women respectively [3]. Asymptomatic bacteriuria is a major predisposing factor for pyelonephritis, a condition linked to obstetrical complications, preterm labor and low infants’ birth weight [3].
UTIs are common causes of maternal and fetal morbidity and mortality [4]. Bacteriuria is predisposed to anatomical and physiological changes imposed on the urinary tract by the hormone progesterone [5]. Progesterone is necessary for implantation of the fertilized egg in the uterus and for maintaining pregnancy [6]. The adjustments driven by progesterone hormones increase in concentration during pregnancy when the maternal immune system is lowered to enable immune tolerance towards paternal antigens expressed on fetal cells [7]. Lowered immunity promotes the growth of pathogens including E. coli in the urinal system of pregnant women.
Increased risk of infection among pregnant women has been attributed to physiological changes, but little emphasis has been made to determine the association between E. coli occurrence and the progesterone hormone levels [5].
The main aim of the current study was to determine the relationship between pregnancy trimesters, changes in progesterone levels and occurrence of UTIs detected in the urine among pregnant women attending antenatal clinic at Moi Teaching and Referral Hospital.
2. Materials and Methods
2.1. Study Area and Study Population
The study was conducted among pregnant women attending antenatal clinic at Moi Teaching and Referral Hospital (MTRH) in Eldoret, which is situated in the mid-west of Kenya’s Rift Valley and shares common borders with Trans Nzoia County to the North, Elgeyo Marakwet County to the East, Baringo County to the South East, Kericho County to the South, Nandi County to the South West and Kakamega County to the North West. Moi teaching and referral hospital is a referral hospital facility with a catchment population from western Kenya. The facility was purposely chosen because of its elaborate clinic for mother to child infection prevention programme, good patient flow and increased urinary tract infections in pregnant women.
2.2. Study Design
A cross sectional design was used and sample size was calculated using the formula of Agresti [8] and Agresti [9] which determined a sample size of 78 participants. All women attending antenatal clinic at MTRH were randomly selected and those who consented and had not used antibiotics during the previous one week were recruited for the study. Every 8th consenting pregnant woman who met inclusion criteria in age groups 15 - 19 (3), 20 - 29 (46), 30 - 39 (25), and 40 - 49 (4) was recruited for the study until the sample size was achieved thus 26 participants for each trimester).
2.3. Sampling Frame
The patients flow in MTRH antenatal clinic for 20 working days is approximately two hundred. The sampling frame of three months was used and 78 samples of both urine and blood were collected by appropriate probability sampling method.
2.4. Sampling Technique
Probability sampling method was used and 78 blood and urine samples were collected at Intervals for 60 working days. Approximately 600 samples are collected at antenatal clinic in 60 working days and 78 samples were required for the study, therefore a sample interval of 8 was used. Every 8th pregnant woman who met the inclusion criteria and consented to participate was sampled until 78 participant were sampled
2.5. Sample Size Determination and Sampling
The sample size required for a logistic regression, in which the predictor is quantitative is given by (Agresti, 2002, 2007).
where
Thus
where λ is Lambda, α Alpha, β Beta δ delta and n population size
This sample size was sufficient to model the relationship between the quantitative predictor x, here progesterone levels, and the outcome, presence of E. coli. The incidence of E. coli from a study conducted at the University of Nairobi was 40% [10]. For simplicity and reasons for computing the sample size, it was assumed that this prevalence value was at an average progesterone level. Higher progesterone levels have been associated with higher incidence of E. coli [10]. Using the incidence of 40% from the study conducted in Nairobi, a sample size that allowed the test to be sensitive to at least a 25% difference in the prevalence of E. coli was required, that is, to increase to 65% from the population prevalence, at one standard deviation increase in the predictor, progesterone levels. The odds of E. coli occurrence at the mean value of progesterone is 0.40/0.60 = 0.67, and the odds of E coli occurrence at a unit standard deviation above the mean are 0.65/0.35 = 1.9. This means that there is odds ratio hence, the chance of type II error in a test. Thus, the sample size required for one predictor is 62. However, for a multivariate logistic regression model where we have other covariates to adjust for in the model, an assumption was made about the correlation between progesterone and the rest of the covariates. It was assumed that R2 is 0.2 for moderate correlation. To get the sample size that was sufficient to study the relationship between the outcome and the anticipated factors division of the sample size was obtained above by (1 − R2), which gave 78 as the number of participants
2.6. Inclusion and Exclusion Criteria
All pregnant women attending antenatal clinic at MTRH who consented to the study were recruited after fulfilling the inclusion criteria. For instance, consenting pregnant women who had not used antibiotics during the previous one week were recruited. However, pregnant women who had used antibiotics during the same period together with those who did not consent to the study were excluded.
2.7. Urine Specimen Collection, Processing and Analysis
Study participants were advised to collect mid stream urine and were provided with a sterile screw capped urine container. About 0.001 ml urine specimens were transferred using calibrated loop onto cysteine lactose electrolyte deficient (CLED) agar media [11] and incubated aerobically at 37˚C for 24 hours. Any microorganism isolated with colony counts of 105 CFU/mƖ of urine was considered positive for urinary tract infection [11]. Bacteria were identified to species level using Gram staining and biochemical tests.
2.8. Blood Specimen Collection and Processing
Blood was collected from each participant by a phlebotomist using standard procedure. Plasma was obtained from 2 mƖ of blood centrifuged at 3000 - 5000 rpm at room temperature and aliquoted into cryovials and used to determine progesterone levels by ELISA technique [12]. The samples were analysed in duplicates and the mean levels were determined and recorded.
2.9. Progesterone Levels Determination
Blood specimens in microtiter plates containing progesterone hormone conjugate were shaken and incubated at room temperature for an hour. Addition of substrate generated hormone-substrate complex with an optimal color after 30 minutes. The extent of color development was inversely proportional to the amount of progesterone in the specimen and was determined by measuring the absorbance (at 450 nm) of the specimens against the standards using a microtiter plate reader (BIOTEK S/N 200136) [12].
2.10. Data Management and Analyses
Data analysis was done using STATA version 13 SE. Categorical variables were summarized as frequencies with corresponding percentages. Continuous variables were summarized by mean and standard deviation (SD). Differences between two normally distributed continuous variables (such as mean progesterone levels during different trimesters) were evaluated using two-sample t-test while those with discontinuous variables (such as E. coli occurrence or non occurrence) were tested by chi-square technique.
3. Results
Out of the 78 urine specimens that were tested, 11 were found to be positives for bacterial pathogens. The most frequent bacterial species that was observed was Escherechia coli (63.7%) and the least was Klebsiella spp (9.1%) of positive samples as shown in Table 1. However, the overall sampled population prevalence of E. coli was 9.3% (Table 3).
The mean progesterone levels at first and second trimesters were within normal ranges; 9 - 47 ng/ML and 17 - 146 ng/ML for first and second trimester respectively (Table 2). However, there was exponential increase of progesterone hormone from trimester one (1) to trimester three (3), with significant differences (p < 0.5) in mean plasma progesterone concentration among pregnant women during the different trimesters. The mean progesterone levels were
![]()
Table 1. Bacterial pathogens identified in the urine of pregnant women attending antenatal clinics at MTRH.
![]()
Table 2. Progesterone concentrations (ng/ML) by trimester in pregnant women attending antenatal clinic at MTRH. Means followed by different letters within a column are significantly different at p < 0.05.
(Normal ranges: 9 - 47 ng/ML, 17 - 146 ng/ML, 43 - 300 ng/ML).
above normal range (43 - 300 ng/ML) during the third trimester as shown in Table 2.
The distribution of E. coli during the three gestation periods (trimesters) is shown in Table 3. There was a higher percentage of E. coli infection during the third trimester (12%) compared to second and first trimesters, which had 8% infections each. However, chi-square statistics showed that the observed differences were not significant (p > 0.05).
The prevalence of E. coli infection in relation to age is shown in Table 4. A majority (58.9%) of participants in this study were pregnant women between age-group 20 - 29 years. However, pregnant women of age group 30 - 39 years had the highest incidence of E. coli infection (16 %) followed by those in age group 20 - 29 years who had 6.5% E. coli infection. There was no E. coli incidence in pregnant women of age groups 15 - 19 years and 40 - 49 years probably due to the fewer numbers of these age-groups among the participants. However, Pearson chi-square test showed insignificant association (p > 0.05) between the age-groups and E. coli occurrence (Table 4).
4. Discussion
Majority of pregnant women who participated in this study were within the child bearing age of between 20 - 39 years which is consistent with the report of Campbell [13] who conducted interviews among senior women, majority of whom reported that pregnancy in early age (≤19 years) is shunned by the society as it undermines the future of young girls. Similarly, fewer pregnant women of 40 years and above were encountered since at this advanced age, women experience infertility and other social and health complications, which deter them from getting pregnant [14].
The most prevalent bacteria isolated in urine of the pregnant women attending ANC at MTRH was Escherechia coli. Pseudomonas aeruginosa and Kleibsiella spp. were infrequently encountered. This could be because uropathogenic E. coli expresses a multitude of virulence factors to break the inertia of the mucosal
![]()
Table 3. Distribution of E. coli frequency during the three gestation periods (trimesters).
![]()
Table 4. Prevalence of E. coli in pregnant women in relation to age.
barrier, and can persist within the urinary tract serving as a reservoir for recurrent infections and serious complications. These results are consistent with those of Zen et al. [15] and Ronald [5] who found that E. coli represents 80.0% of bacterial isolates in bacteriuria while the current study E. coli result was at 63.7% of bacterial isolates in bacteriuria. However, Tadesse et al. [16] observed a lower prevalence rate (10.4%) of E. coli which is comparable to our result of 9.3% E. coli prevalence among all pregnant women in the study. Pseudomonas spp. was isolated in only one case in the present study. The very low growth of Pseudomonas spp. could be attributed to the fact that all the investigated participants were from the outpatient department and Pseudomonas spp is more commonly acquired as a nosocomial infection following prolonged hospital stay. Hamdan et al. [17] and Ronald [5] also reported similar low prevalence results for this species.
There was an exponential increase in progesterone hormone among study participants from trimesters one to trimester three, which is consistent with the observation that progesterone levels increase as pregnancy progresses [3]. The increase was progressive from the first trimester, reaching a peak during the third trimester of pregnancy as demonstrated by our data (Table 2). Progesterone is primarily produced by the placenta during pregnancy [4]. Progesterone levels during pre-ovulatory phase of the menstrual cycle are relatively low, and it rises after ovulation and is elevated during the luteal phase [10]. If pregnancy occurs, human chorionic gonadotropin is released maintaining the corpus luteum, allowing it to maintain appropriate levels of progesterone [4].
The results of the current study show that the frequency of E. coli was higher during the third trimester compared to the first and second trimesters. This probably occurred because of weakening of immune systems of the body as pregnancy progresses. Jennifer et al. [18] reported a similar occurrence and observed that changes due to pregnancy results in increase in bladder volume while detrusor tone decreases, and leads to relaxation of ureteric smooth muscles causing ureter dilatation, which could lead to urinary stasis hence supporting the colonization of the bladder by bacterial organisms. Onoh et al. [19] and Emiru et al. [20] also found an increased frequency of E. coli during the third trimester compared to the first and second trimesters. This may also be as a result of the pressure effect of a bigger uterus on the ureter at the third trimester, and also the increasing smooth muscle relaxing effect of pregnancy hormones and the pressure on the bladder which may lead to stasis of urine which encourage bacteria multiplication. Onoh et al. [19] also reported a higher prevalence of E. coli infection in the second trimester compared to the third trimester. However, this observation needs further research to confirm. Furthermore, several individual and socio-economic factors can impact on exposure and susceptibility to infection. The possible confounding factors also need to be evaluated.
A majority of participating pregnant women were between age-group 20 - 29 years. However, there were fewer participants between age-groups 15 - 19 years and 40 - 49 years respectively probably because the former age-group comprise of girls who are pursuing formal education and are less likely to get involved in sexual activities or are probably using safe sex practices while the latter consists of women who have probably obtained the desired number of children, and as a result, are practicing family planning to prevent unwanted pregnancies as suggested by the KDHS [21] data.
High prevalence of E. coli in pregnant women between 30 - 39 years was reported in the present study. The reason could be due to the fact that many women within this age group are likely to have had many children before the present pregnancy and it has been reported that multigravida is a risk factor for acquiring E. coli in pregnancy [15]. However, certain contraceptive methods are also said to increase the risk of contracting an infection, although women are most sexually active at this age [20]. Furthermore, this could be due to lower pH compared to the genital track of younger women whose genital tracks are acidic, which is inimical to the growth of bacterial organisms including E. coli. However, the scope of our protocol was limited to the selected pregnant outpatient participants whose physiological conditions were not investigated.
The current study showed that progesterone levels significantly increased during third trimester and the most prevalent bacteria associated with this increase was E. coli even though age and increase in progesterone levels had no significant association with E. coli infection. Further research is required to identify bacterial infections among the neonates in hospitals to determine the transmission dynamics of these infections. The study by Barcaite et al. [22] found that mothers who had E. coli bacterial infections were also infected by group B Streptococcus, which is a major cause of fatal neonatal sepsis, although these deaths could easily be controlled by screening and treatment of infected women during pregnancy. It is therefore recommended that pregnant women should routinely be diagnosed and treated of uropathogens to improve maternal and neonate health.
Abbreviation
UTI Urinal Tract infection;
CFU Colony Forming Unit;
CLED Cystine lactose Electrolyte Deficient Agar;
MTRH Moi Teaching and Referral Hospital;
ELISA Enzyme-Linked ImmunoSorbent Assay;
Spp Species.