Prevalence of urinary tract infection and the validity of urine analysis in the diagnosis of uti in febrile under-fives at Nnamdi Azikiwe University Teaching Hospital, Nnewi: a cross-sectional study

Abstract

Background

Urinary tract infection (UTI) is a common cause of fever in febrile under-fives, but this is often overlooked due to the ambiguity of its presentation, especially in this malaria-endemic region. This makes clinical recognition difficult despite the potential risk of renal scarring and chronic kidney disease (CKD) later in life, if inappropriately treated. Additionally, urine culture, the gold standard for diagnosis, is limited by its time-consuming nature. Due to the wide variation in reports on the accuracy of the dipstick test as a rapid screening test for UTI in children, studies are necessary to validate its relevance in the diagnosis of UTI in different practice localities. This study evaluated the effectiveness of dipstick urinalysis in diagnosing UTIs in febrile children under five years old at Nnamdi Azikiwe University Teaching Hospital (NAUTH), Nnewi.

Methods

Three hundred thirty children aged 1 month to 59 months with an axillary temperature of ≥ 37.5 °C were recruited consecutively from July 2016 to April 2017 from the Children Outpatient (CHOP) and the Children Emergency Room (CHER). Sociodemographic and clinical data were collected using interviewer-administered questionnaires. Urine samples were collected from each participant for dipstick urinalysis, microscopy, and culture testing. Data were analysed using SPSS software version 21. The level of significance for tests of association was set at 5%.

Results

There were 218 males and 112 females, resulting in a male-to-female ratio of 1.9:1. Subjects aged 1 to 11 months had the highest frequency, comprising 31.5% of the study population. The prevalence of UTI based on urine culture was 4.8% and it was significantly higher in females than in males. Temperature ≥ 38.4 °C had a significant association with the presence of UTI. No other clinical feature was significantly associated with UTI. Using urine culture as the gold standard, the sensitivities of Leucocyte esterase (LE), nitrite, and their combination (LE and/or nitrite) were 31.3%, 25% and 37.5% respectively, while their specificities were 97.1%, 100% and 97.1% respectively. Gram stain of uncentrifuged urine had the best accuracy of the individual tests studied. E. coli was the commonest pathogen, which was isolated in 43.8% of children with UTI, while meropenem, cefepime, and ceftazidime had the highest antimicrobial activity.

Conclusion

UTI is common in febrile under-fives, and dipstick urinalysis is a good screening test, recommended for routine use in this population. E. coli is the commonest organism responsible for UTI in febrile under-fives in NAUTH, Nnewi, followed by Klebsiella spp. All isolated organisms were sensitive to Meropenem, while Cefepime and Ceftazidime also had high antimicrobial activities and should be used in the empirical treatment of the disease.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-12617-y.

Introduction

Fever is a common reason for hospital visits in children under five [1, 2], especially in sub-Saharan Africa, where infectious causes like malaria are predominant [3–6]. Due to malaria’s endemicity, other causes like urinary tract infections (UTIs) are often overlooked. UTIs are common in young children and can coexist with malaria, but diagnosis is challenging because symptoms are non-specific, clinical recognition is difficult, and collection methods may be unreliable or distressing [7–15]. It is probably the most common disease of the urogenital tract in childhood [6], particularly in infants and children younger than two years [17].

Urine culture is the gold standard, [18, 19] but takes time, requires expertise, and is costly, so rapid tests like dipstick urinalysis and microscopy are valuable for prompt screening [20, 21]. Dipstick tests for leukocyte esterase and nitrites are practical, affordable, and quick, helping identify likely UTIs, though they cannot replace culture. Although microscopy can be performed immediately, it is not always reported until the culture result is available. Furthermore, it generally demands expertise that is not widely available and also requires quality assurance, which can be a drawback [19]. The dipstick strip contains test pads for leukocyte esterase (LE), nitrite, blood, protein, glucose, ketones, bilirubin, urobilinogen, pH, and specific gravity in various combinations. Although the application of some of these components of a dipstick, which include blood, protein, and glucose, in the diagnosis of UTI has been studied [22], LE and nitrite, which signify pyuria and bacteriuria, respectively, are the components basically used in clinical practice with regard to the diagnosis of UTI [21]. These were the test elements of dipstick urinalysis used in this research. A wide range of dipstick accuracy has been reported from studies, which is influenced by various factors such as age and symptoms of the study population, method of urine collection, and inherent accuracy of the test strip [21, 22].

Data from the Urologic Disease in America Project show that pediatric UTI is a significant healthcare burden, affecting 2.4% to 2.8% of children annually and causing over 1.1 million doctor visits and 5% to 14% of emergency room visits. Inpatient costs exceed $180 million per year, with total expenses likely much higher [23]. Similar data is lacking for Nigeria, but with a prevalence of 9% to 13.7%, pediatric UTI likely also places a substantial strain on the country’s healthcare system. UTI in infants and young children is often linked to urinary tract abnormalities, with vesicoureteric reflux (VUR) present in 35–40% of first-time UTI cases [24–26]. Most febrile children with UTI have acute pyelonephritis, warranting management for this condition [27, 28]. About 50% develop renal scarring [18], which occurs as renal defects in 5–15% within 1–2 years, [29] especially in children under two [30]. Persistent kidney damage increases the risk of long-term issues like impaired renal function, hypertension, and ESRD [31, 32]. Prompt diagnosis and treatment of febrile UTI are crucial to prevent complications and identify underlying abnormalities, particularly in developing countries like Nigeria, where CKD often results from acquired conditions [33, 34].

Due to the increasing prevalence of CKD [34, 35] and its global burden, inadequate facilities for the management, widely unaffordable services where available, and with several aetiologies beginning in childhood, it is paramount that efforts should be channelled towards preventive measures to reduce the effect of CKD in society. Early diagnosis and appropriate treatment of UTI in febrile children is one of such measures [36]. In line with this, we sought to determine the prevalence and clinical presentations of UTI in febrile under-fives; the accuracy of dipstick for leukocyte esterase and nitrite tests, the value of microscopy for pyuria and bacteriuria in the diagnosis of UTI using urine culture as gold standard; and the pathogens responsible for UTI in febrile under-fives and their antibiotic sensitivity patterns.

Methodology

Study site

This study was conducted at the Paediatric unit of the Nnamdi Azikiwe University Teaching Hospital (NAUTH), a 294-bed tertiary hospital [37]. It is situated in Nnewi, which is the biggest industrial city and amongst the major urban cities in Anambra State, South East Nigeria [38] and is accessed by patients from all over the state and nearby towns of neighbouring states.

Study design

This was a cross-sectional study conducted from July 2016 to April 2017.

Study population

All febrile paediatric patients aged 1 to 59 months who presented at the Children’s Outpatient (CHOP) clinic, paediatric specialist clinics, and the Children’s Emergency Room (CHER).

Inclusion criteria:

• i.Children aged 1 to 59 months.
• iii.Presence of fever (axillary temperature ≥ 37.5 °C) [39], with or without suspected focus of infection.
• iii.Children whose parents/caregivers gave written informed consent.

Exclusion criteria:

Patients with any of the following were excluded from the study;

• i.History of use of antibiotics or antimalarials that have antibacterial activity, such as sulfadoxine/pyrimethamine, within 14 days before enrolment into the study.
• ii.Urologic manipulations such as an indwelling urethral catheterisation, carried out within 48 h preceding the evaluation [40].
• iii.Known history of kidney or urinary tract structural abnormalities as determined from their medical records, or those whose symptoms suggested abnormality of the urinary tract, such as poor urinary stream.
• iv.Prior definitive diagnosis of medical conditions associated with immunosuppression, such as severe malnutrition, sickle-cell anaemia, malignancies, human immunodeficiency virus infection/acquired immunodeficiency syndrome, or subjects on immunosuppressive drugs,
• v.Prior definitive diagnosis of co-existing renal pathologies like nephrotic syndrome, glomerulonephritis, and chronic renal failure.

Sample size

Sample size was calculated using Buderer’s formulas for sensitivity (Nse) and specificity (Nsp) of a diagnostic test study, with the final minimum sample size (N) being the larger number between Nse and Nsp [41].

Data collection

Structured questionnaires (Appendix 1), which were interviewer-based, were used by the investigator to collect relevant information, such as sociodemographic and clinical data from the parents/caregivers. Socioeconomic class was assigned to each participant using Oyedeji’s classification [42]. A thorough physical examination was carried out on each participant by the investigator to identify signs suggestive of UTI. A clinical diagnosis of UTI was made in subjects with any of the following: costo-vertebral angle tenderness, crying or pain on micturition, suprapubic tenderness, increased urinary frequency, loin pain and/or tenderness [19].

Specimen collection

A sample of mid-stream urine was collected from toilet-trained participants mostly above two years with the help of their parents or caregivers. In females, the parents/caregivers parted their labia during micturition, and in all subjects, the first portion of the urine was allowed to fall outside the container, and then about 10 ml of midstream urine was collected. It was ensured that the specimen bottle did not come in contact with the perineum, external genitalia, or adjacent skin to avoid contamination. For infants and younger children who were not toilet trained, 5 mL of urine was collected by the investigator using supra pubic aspiration (SPA) under aseptic conditions after it was established that the participant had not passed urine for at least one hour before the procedure (duration on dry diaper or from last observed urination). If the elapsed time since the last void was less than one hour or SPA was unsuccessful, in-and-out urethral catheterisation under aseptic conditions was done. The initial drops of urine were not collected into the specimen bottle.

When SPA or catheterisation was to be performed, a research assistant had a sterile container ready to collect a clean catch urine specimen because the preparation for the procedure could stimulate the child to void. All specimens were collected in labelled sterile containers and sent to the microbiology laboratory of the hospital within one hour of collection, and were processed immediately by the laboratory collaborator. In a situation where a specimen could not be processed for culture within an hour of collection, one-half of the sample was used for dipstick urinalysis, while the other half was refrigerated at 4 °C for less than 24 h from the time of collection prior to microscopy and culture [43]. Potentially contaminated samples were discarded, and the patients resampled.

Test methods

After inoculation of the fresh urine specimen on culture plates for culture, urine microscopy and dipstick urinalysis were then carried out on the rest of the sample.

Dipstick urinalysis

This was carried out using Combur 10 test strips made by Roche Diagnostics GmbH, Germany, according to the manufacturer’s standard operating instructions. It has a practical detection limit of 10 to 25 leukocytes per microlitre for the LE test and 0.05 mg/dL for the nitrite test [44]. A strip was dipped into each participant’s urine for one second, ensuring that all test pads were fully immersed, after which excess urine was tapped off. After 60 s, the resultant colour on the test area of the strip was compared against the respective colour scale on the test container under good light, and the result was recorded. For the nitrite test, any degree of uniform pink colour was interpreted as positive, while a change from beige to violet or purple on the LE pad was reported as positive as +, ++, or +++ based on the matched colour scale on the dipstick container. Participants without clinical evidence of UTI who had nitrite-positive or both nitrite and LE-positive tests were presumed to have UTI based on dipstick result, with definitive diagnosis dependent on culture results [19]. The dipstick results were communicated to the parents/guardians immediately and presented to the attending physicians.

Microscopy

A direct wet preparation of a drop of uncentrifuged urine was viewed under the light microscope (Olympus) for WBC, red blood cells (RBC), bacterial debris, motility, epithelial cells, casts, and crystals. Viewing was done using a low-power x10 objective initially and then a high-power x40 objective later. The presence of pyuria of ≥ 2 WBCs/HPF was regarded as significant and suggestive of UTI [47]. This cut-off count is the same for urine obtained from suprapubic aspiration and urethral catheterisation [21, 45].

Gram staining of uncentrifuged urine was done as described in the Manual of Clinical Microbiology [46]. A circle of about 1.5 cm in diameter was drawn on the underside of a clean, grease-free labelled slide using a marker to delineate the area in which the smear was made. Using a sterile loop, two drops of urine were placed within the encircled area and spread using a circular motion to form a thin smear about one cm in diameter. This was allowed to air dry (about 30 min), and heat-fixed, bypassing the slide through the flame of a Bunsen burner two or three times with the smear side up. This heat fixation denatures bacterial proteins, causing the cells to stick to the slide and take up stains more readily. The heat-fixed slide was placed on a staining tray, the smear covered with crystal violet (primary stain), and allowed to stand for one minute, after which the slide was tilted and the stain rinsed off with distilled water using a wash bottle. The smear was covered with Gram’s iodine for one minute and rinsed off with distilled water. Holding the slide at a 45 ° angle, a decolouriser was carefully added one drop at a time until the solution drained was no longer coloured. At this point, the slide was immediately rinsed off with distilled water. The smear area was finally covered with safranin (counterstain) for 45 s, rinsed off with distilled water, blotted with bibulous paper, and viewed under oil immersion using a light microscope. Any bacteria per oil immersion field was considered significant. Subsequently, 5 ml of urine was centrifuged in a test tube at 3,000 revolutions per minute (rpm) for five minutes, after which a wet preparation was made with one drop of the sediment. Pyuria of > 5 WBCs/HPF or at least 10 bacteria per HPF was considered significant [47]. Appropriate collection of midstream urine, especially in females, to avoid contamination with vaginal secretion, ensures this cut-off for pyuria is the same for urine obtained midstream, from suprapubic aspiration and urethral catheterisation [21, 45]. Participants without clinical evidence of UTI who had bacteriuria and pyuria positive or only bacteriuria positive were presumed to have UTI based on microscopy [19].

Culture and sensitivity

Culture of all urine samples was done. Each uncentrifuged urine specimen was well-mixed and inoculated on plates of MacConkey and chocolate agar, as also described in the Manual of Clinical Microbiology [46], using a standard calibrated sterile wire loop which delivers 0.001 ml of urine per loopful. Thereafter, the plates were incubated aerobically at 37 °C for 24 h, after which the colonies were counted. Any growth on culture of urine obtained through suprapubic aspiration, growth of greater than 10,000 CFU/ml of a single pathogen on catheter specimen urine or growth of more than 100,000 CFU/ml of a single pathogen on clean voided urine (≥ 10⁴ CFU/ml in a child with symptoms of UTI) was regarded as significant bacteriuria and therefore diagnostic of UTI [14, 18, 47, 48]. Growth less than the appropriate diagnostic criteria and mixed organisms generally indicate contamination [48].

The AAP clinical practice guideline for the diagnosis and management of initial UTI in febrile infants and children 2 to 24 months recommended that to establish the diagnosis of UTI, clinicians require both urinalysis results that suggest infection (pyuria and/or bacteriuria) and the presence of at least 50,000 CFU/ml of a uropathogen cultured from a urine specimen obtained through catheterization or SPA [21]. However, this recommendation, which is limited to infants and children less than two years, is not yet widely accepted and practised [48]. Thus, many centres still use the diagnostic criteria stated initially for significant bacteriuria. In cases with significant growth, systematic bacteriologic and biochemical tests using standard techniques were carried out to identify the bacterial species. Growth less than the appropriate diagnostic criteria and mixed organisms generally indicates contamination [43, 48]. Antimicrobial susceptibility pattern of isolated pathogens was determined by the Kirby-Bauer diffusion method using single antibiotic discs by Oxoid. The diameter of the resultant zone of growth inhibition was measured and reported as sensitive, intermediate, or resistant to the drugs that were tested. Drugs with sensitive or intermediate results were considered effective against the respective organisms cultured. The selection of antibiotics tested and the interpretation of their zones of inhibition were based on the Clinical Laboratory Standard Institute (CLSI) guideline and interpretative chart for antimicrobial susceptibility testing. Standard quality control was observed in all steps of the tests. Results of the tests were entered into each participant’s proforma and were also communicated to the caregivers.

Treatment

Participants with a presumed diagnosis of UTI made on clinical grounds and/or urinalysis were commenced on empirical treatment by the managing unit using oral amoxicillin/clavulanic acid or cefixime for outpatients, while inpatients were given intravenous ceftriaxone or ceftazidime [21, 25]. Treatment was subsequently modified in line with culture and antibiotic sensitivity results. Subjects with culture-confirmed UTI were referred to the nephrology unit for further review, treatment, and follow-up.

Data entry and analysis

Data were analysed using the Statistical Package for the Social Sciences (SPSS) software version 21 for Windows. Descriptive statistics were used to describe the mean, median, and standard deviation of the continuous variables, such as age, duration of fever, and axillary temperature on presentation, while frequencies were used for categorical variables. Age, axillary temperature, and duration of fever were subsequently categorised for further analyses. Decision tree analysis on SPSS was used to identify the cut-off value of axillary temperature, which was significantly associated with UTI. Categorical variables, namely sex, symptoms, and signs, were tested for association with UTI using Pearson’s Chi-square (and Fisher’s exact test where appropriate) for bivariate analysis, and logistic regression was used for multivariate analysis. Student’s t-test was used to compare the difference in means of continuous variables like axillary temperature between subjects with UTI and those without UTI. The level of significance was set at 5%.

Results

A total of 330 children were included in the study. Urine specimen was collected by SPA in 61 children, by urethral catheterisation in 102, mid-stream in 162, and clean catch in 5 subjects. Of the 330 participants recruited, there were 218 (66.1%) males (All circumcised) and 112 (33.9%) females, giving a male-to-female ratio of 1.9:1. Infants constituted 31.5% of the enrolled participants. The majority of the participants were from the middle Sect. (45.1%) (Table 1). The mean age was 26.37 ± 18.68 months, the median age was 22 months, and the modal age was six months, which implies that the age of enrolled subjects had a right-skewed (positive)distribution.

Table 1.

Sociodemographic characteristics of study subjects

Characteristics	Frequency (%) N = 330
Sex
Males Females	218 (66.1) 112 (33.9)
Age group
1–11 months 12–59 months	104 (31.5%) 226 (68.5%)
Socioeconomic class
Upper class Middle class Lower class	93 (28.2) 149 (45.1) 88 (26.7)

The mean axillary temperature of the population studied was 38.5 ± 0.6 °C (range 37.5 to 40.5 °C) while the mean duration of fever was 3.6 ± 3.5 days (range < 1 to 21 days). The common associated symptoms obtained in the study population were cough, which was present in 121 subjects (36.7%), poor feeding (26.7%), weakness (21.2%), vomiting (20.3%), catarrh (19.7%), diarrhoea (11.5%) and abdominal pain (9.4%). There were 37 subjects (11.2%) with fever as the only symptom, three subjects had suprapubic pain (0.9%), two had dysuria (0.6%), one had increased urinary frequency (0.3%), while bloody urine was present in one child (0.3%). There were 256 children without systemic findings on physical examination. The most frequent system with signs suggestive of underlying disease and, by extension, cause of fever was the respiratory system, with findings obtained in 27 subjects (8.2%). Only one child had a clinical sign suggestive of UTI, which was suprapubic tenderness (0.3%). Out of the 330 patients studied, 16 children had a UTI diagnosed by significant bacteriuria on urine culture, giving an overall prevalence of 4.8%. No mixed growth or colony count less than the appropriate diagnostic criteria was obtained on culture.

The prevalence of UTI in females (9.8%) was significantly higher than in males (2.3%) {p < 0.05}, especially in children 12 to 59 months. There was no significant difference in the prevalence of UTI between infants (7.7%) and children 12 to 59 months (3.5%) {p > 0.05}. However, for male subjects with UTI, prevalence was significantly higher in infants (6.5%) compared with those 12 to 59 months (0.6%) {p < 0.05 (Fisher’s exact)} (Table 2).

Table 2.

Differences in the prevalence of UTI by gender and age group

UTI Characteristics	Frequency	UTI Present (%)	Χ2	P value
All females All males	112 218	11 (9.8) 5 (2.3)	9.08	0.003*
Females 1 − 11 months Males 1 − 11 months	42 62	4 (9.5) 4 (6.5)	-	0.712†
Females 12–59 months Males 12–59 months	70 156	7 (10.0) 1 (0.6)	-	0.001*†
Subjects 1–11 months Subjects 12–59months	104 226	8 (7.7) 8 (3.5)	2.66	0.103
Males 1 − 11 months Males 12 − 59 months	62 156	4 (6.5) 1 (0.6)	-	0.024*†
Females 1 − 11 months Females 12–59 months	42 70	4 (9.5) 7 (10.0)	-	1.000†

† = P value is for Fisher’s exact test. * = statistically significant

There was significant difference in mean temperature between children with UTI (38.8 ± 0.6 °C) and those without UTI (38.5 ± 0.7 °C) {t-value = 2.186, p = 0.03} while there was no significant difference in the mean duration of fever between children with UTI (2.8 ± 1.8 days) and those without UTI (3.7 ± 3.5 days) {t-value = 0.936, p = 0.350}. Furthermore, axillary temperature reading ≥ 38.4 °C was found to be significantly associated with increased prevalence of UTI {p < 0.05}. Although 15 children with UTI had a fever duration of < 7 days, there was no significant difference in the occurrence of UTI between subjects with a fever duration < 7days and those ≥ 7 days (Table 3).

Table 3.

Relationship between fever and UTI

Variable	No. of subjects	UTI present (%)	UTI absent (%)	Χ2	P value
Axillary temperature‡
< 38.4 °C ≥ 38.4 °C	164 166	2 (1.2) 14 (8.4)	162 (98.2) 152 (91.6)	9.307	0.002*
Duration of fever
< 7 days ≥ 7 days	284 46	15 (5.3) 1 (2.2)	269 (94.7) 45 (97.8)	-	0.709†

‡ = significant cut-off value was detected by use of a decision tree on SPSS, † = P value is for Fisher’s exact test. * = statistically significant

Fever alone or fever with any other symptom was not significantly associated with the presence of UTI in the study population (Table 4). There were three subjects with symptoms related to the urinary system, with suprapubic pain present in all three. Two out of these three subjects had dysuria, while increased frequency of urination was present in one. Of the 127 subjects with cough and/or catarrh, none had significant bacteriuria on urine culture. This relationship between the presence of cough or catarrh and the absence of UTI in this study was found to be significant (p values of 0.002 (χ2 = 9.735) and 0.049 (Fisher’s exact), respectively).

Table 4.

Presenting symptoms and occurrence of UTI

Symptom‡	Frequency	UTI present	UTI absent	χ2	P value
Poor feeding	88	4	84	-	1.000†
Weakness	70	5	65	-	0.346†
Vomiting	67	5	62	-	0.334†
Diarrhoea	38	4	34	-	0.098†
Fever alone	37	4	33	-	0.091†
Abdominal pain	31	1	30	-	1.000†
Suprapubic pain	3	1	2	-	0.139†
Dysuria	2	1	1	-	0.095†
Urinary frequency	1	0	1	-	-
Haematuria	1	0	1	-	-

‡ = Some subjects had more than one symptom † = P value is for Fisher’s exact test

Malaria was the most prevalent provisional diagnosis in the study (36.7%), followed by URTI (25.6%). For the 16 subjects with confirmed UTI, malaria was the provisional diagnosis in six (37.5%), acute gastroenteritis in three (18.8%), sepsis in three (18.8%), fever without localising signs in two (12.5%), UTI in one (6.2%), and impetigo in one (6.2%). Provisional diagnosis had no significant association with UTI.

On multivariate analysis, significant factors with higher odds for UTI were female sex {OR = 4.449, p < 0.05} and presenting axillary temperature ≥ 38.4 °C {OR = 7.190, p < 0.05}. This means that febrile female children less than five years are approximately four times more likely to have UTI than their male counterparts, whereas febrile under-fives with axillary temperature ≥ 38.4 °C are about seven times more likely to have UTI than those with a temperature below 38.4 °C.

On macroscopic examination of urine, 13 subjects (3.9%) had turbid urine with UTI present in six of them. Turbid urine was significantly associated with UTI {p < 0.001 (Fisher’s exact)}.

Using urine culture as the gold standard, the sensitivity of the positive (+) leukocyte esterase test was 31.3% with a specificity of 97.1%. Of the three subjects with a clinical diagnosis of UTI, two had a positive LE result. The nitrite test had a sensitivity of 25% and a specificity of 100%. Table 5 shows the results of the tested parameters of dipstick urinalysis, while their measures of accuracy are shown in Table 6.

Table 5.

Results of dipstick Urinalysis using urine culture as gold standard for diagnosis of UTI

Test Parameter	Result	UTI Present	UTI absent	Total
LE only	Positive Negative Total	5 11 16	9 305 314	14 316 330
Nitrite only	Positive Negative Total	4 12 16	0 314 314	4 326 330
LE and/or Nitrite	Positive Negative Total	6 10 16	9 305 314	15 315 330

LE = Leukocyte Esterase

Table 6.

Different measures of accuracy of dipstick Urinalysis using urine culture as gold standard for UTI diagnosis

Test	SN (%)	SP (%)	PPV(%)	NPV (%)	FPR (%)	FNR (%)	DA (%)
LE only	31.3	97.1	35.7	96.5	2.9	68.7	93.9
Nitrate	25	100	100	96.3	0	75	96.4
LE and/or Nitrite	37.5	97.1	40	96.8	2.9	62.5	94.2

SN = Sensitivity; SP = Specificity; PPV = Positive predictive value; NPV = Negative predictive value; FPR = False positive rate; FNR = False negative rate; DA = Diagnostic accuracy, hpf = high power field

Microscopic examination of uncentrifuged urine for pyuria had a sensitivity of 62.5% and a specificity of 95.5% whereas pyuria on centrifuged urine had a sensitivity of 56.3% and a specificity of 96.8%. Gram stain of uncentrifuged urine had the best harmony of sensitivity (93.8%), specificity (99.7%), positive predictive value (93.8%) and negative predictive value (99.7%), and also had the best diagnostic accuracy (99.4%).

Escherichia coli was the commonest uropathogen responsible for UTI in this study and was isolated in 7 of the 16 children (43.8%) with UTI. It was closely followed by Klebsiella pneumoniae isolated in five subjects (31.3%). Other organisms isolated were Proteus mirabilis (2/16, 12.5%), Staphylococcus aureus (1/16, 6.2%), and Klebsiella oxytoca(1/16, 6.2%). The distribution of the isolated pathogens by gender of the subjects with UTI is shown in Table 7.

Table 7.

Distribution of cultured organisms by gender of study subjects with UTI

Isolated organism	Male	Female	Total N (%)
Escherichia Coli	3	4	7 (43.8)
Klebsiella Pneumoniae	1	4	5 (31.3)
Proteus Mirabilis	1	1	2 (12.5)
Klebsiella Oxytoca	0	1	1 (6.2)
Staphylococcus Aureus	0	1	1 (6.2)
TOTAL	5	11	16 (100.0)

All organisms isolated were sensitive to meropenem (100%) and also showed high sensitivity to cefepime (93.8%), ceftazidime (87.5%), and cefoxitin (87.5%). Sensitivities of the pathogens to ampicillin (6.3%), cefuroxime (6.3%), and cefazolin (6.3%) were very poor (Table 8).

Table 8.

Antibiotic susceptibility pattern of isolated pathogens number of sensitive organisms - n (%)

Antibiotics tested	E. coli (n = 7)	Klebsiella Pneumonia (n = 5)	Proteus spp. (n = 2)	Klebsilla Oxytoca (n = 1)	Staph. Aureus (n = 1)	Overall Sensitivity N (%) N = 16
Meropenem	7 (100)	5 (100)	2 (100)	1 (100)	1 (100)	16 (100)
Cefepime	7 (100)	4 (80)	2 (100)	1 (100)	1 (100)	15 (93.8)
Ceftazidime	6 (85.7)	4 (80)	2 (100)	1 (100)	1 (100)	14 (87.5)
Cefoxitin	7 (100)	4 (80)	2 (100)	1 (100)	1 (100)	14 (87.5)
Ceftriaxone	4 (57.1)	2 (40)	2 (100)	0 (0)	1 (100)	9 (56.3)
Ciprofloxacin	5 (71.4)	2 (40)	2 (100)	0 (0)	1 (100)	9 (56.3)
Amoxyclav	3 (42.9)	3 (60)	0 (0)	0 (0)	0 (0)	7 (43.8)
Gentamicin	3 (42.9)	1 (20)	1 (50)	1 (100)	0 (0)	6 (37.5)
Cotrimoxazole	1 (14.3)	1 (20)	0 (0)	0 (0)	1 (100)	2 (12.5)
Cefuroxime	1 (14.3)	0 (0)	0 (0)	0 (0)	0 (0)	1 (6.3)
Ampicillin	1 (14.3)	0 (0)	0 (0)	0 (0)	0 (0)	1 (6.3)
Cefazolin	1 (14.3)	0 (0)	0 (0)	0 (0)	0 (0)	1 (6.3)

Amoxyclav = Amoxicillin Clavulanate

Discussion

The prevalence of UTI in febrile under-fives in NAUTH is 4.8%, despite only 3 of 330 children being clinically diagnosed, a factor noted as unreliable by Kaufman et al.,[49] highlighting the risk of missed or inappropriate treatment. This rate is close to the 3.5–5% reported in Indian studies [50, 51] but lower than rates from Benin [9] (9%), Enugu [7] (11%), and Maiduguri [8] (13.7%). Higher prevalence in some centres may be due to recruitment from emergency units where sicker children present [52]. This study also found higher UTI rates in CHER than CHOP (6.5% vs. 4.3%), as well as in a Kenyan study [53]. This is because most patients admitted to CHER have serious illnesses, including febrile UTIs, whereas a large proportion of febrile children with common illnesses like uncomplicated malaria and upper respiratory infections are reviewed in clinics as outpatients.

Variations in age and sex distribution also explain differences between studies, as female gender and age above infancy are established risk factors. In the study by Ibeneme et al. [7] in Enugu, they had a lower male-to-female ratio (1.3: 1 vs. 1.9: 1) and a greater percentage of children 12 to 59 months (81% vs. 68.5%) when compared with this study. With the female gender known to be a risk factor for UTI typically above infancy, [18] the higher proportion of females and subjects 12 to 59 months in their study could be partly responsible for the higher prevalence they obtained. A U.S study [54] reported a much lower prevalence (1.7%), consistent with fewer infectious diseases in developed countries [55].

Female sex was significantly associated with UTI (9.8% vs. 2.3% in males; OR 4.4), implying that febrile female under-fives are approximately four times more likely to have UTI than their male counterparts, consistent with global data, [7, 56, 57] and this is due to the proximity of their short urethra to the anus which makes it easier for bacteria from the perineum to ascend the urinary tract. No sex difference was found in infants, which is in agreement with the study from Enugu [7] but at variance with reports from some standard paediatric textbooks [18, 47] and publications [26, 52], stating that the prevalence of UTI in infants is higher in males than females due to the apparent contribution of uncircumcision (a known risk factor) and higher occurrence of abnormalities of the urinary tract in them [58]. The exclusion of children with suspected or known structural abnormalities of the urinary tract from this study and the recruitment of males that were all circumcised due to the high practice of circumcision in the region, perhaps reduced the prevalence in male infants.

The likelihood of UTI in febrile children has been documented to decrease with increasing age, with the highest prevalence in infants because of the immaturity of their immune system [59, 60] an assertion corroborated by Ibeneme et al. [7] in Enugu. Although the prevalence of UTI was higher in infants (7.7%) compared to subjects 12 to 59 months (3.5%) in this study, the difference was not significant. This is similar to the finding by Fredrick et al. [61] in Tanzania. Higher fever was strongly associated with UTI: children presenting with temperatures ≥ 38.4 °C were seven times more likely to have UTI. This finding of association between degree of temperature and UTI agrees with findings by Musa-Aisien [9] and the American Academy of Pediatricians [21], suggesting that degree of temperature may be cautiously employed in selecting febrile under-fives that may need further evaluation for UTI, particularly in resource-poor settings where tests for UTI may be inadequate or unavailable. In contrast, Rabasa et al. [8] did not find any association between temperature at presentation and UTI in their study. No clinical feature (including those suggestive of UTI) was significantly associated with UTI in this study. This is in keeping with the finding by Rabasa et al. [8] However, Musa-Aisien et al. [9] reported that the presence of abdominal pain was significantly associated with increased risk of UTI. These different findings could be due to the higher prevalence of UTI in infants who cannot verbalise compared to older children in the index study. Symptoms and signs of urinary tract pathology, characterised by pain such as loin pain and tenderness, are vague in young children who have difficulty communicating. Also, dysuria and suprapubic pain can be caused by chemical irritation (from soaps and skin lotions) [62] and constipation [63], respectively, amongst others. This is demonstrated by the provisional diagnosis of UTI in as few as 3 out of the 330 subjects (0.9%) evaluated, despite the prevalence of 4.8% obtained in the study using the gold standard for diagnosis. Additionally, of the three children with a provisional diagnosis of UTI, only one child had UTI confirmed by urine culture. This means that about 15 children with febrile UTI would have had a missed diagnosis and left untreated, which is unacceptable given its link to CKD, whereas two children without UTI would have been wrongly treated if the management decision was based entirely on clinical diagnosis. Children with respiratory infections uniformly tested negative for UTI, consistent with evidence that an obvious alternative fever source reduces UTI probability, [15, 21] while in an Indian study, [51], they noted that the presence of another potential source of fever, such as upper respiratory tract infection or otitis media, is not reliable in excluding UTI.

Turbidity of urine on macroscopic examination had a significant association with the presence of UTI in this study which is as reported by the University of York [64] and is due to the presence of cells such as WBC, bacteria, and epithelial cells in the urine. For this reason, a clinician assessing a febrile child may need to examine the child’s urine macroscopically, where possible, particularly when further evaluation for UTI is deemed unnecessary.

Detecting UTI by a positive leukocyte esterase test in this study had low sensitivity (31.3%) with a high specificity of 97.1% indicating that a positive result is good for ruling in UTI, while the test is poor in ruling out UTI because 68.7% of children with UTI could have a negative result. Although the LE test generally has a higher sensitivity than specificity [21], some studies [65, 66] have also reported a lower sensitivity than specificity, as observed in this study. The heterogeneity of LE results in these studies may be due to the varied inherent sensitivity of the various test strips used and differences in the population studied. The LE test had a low PPV of 35.7% and a high NPV of 96.5%. This low PPV, which means that the probability that a child with a positive test result has UTI is 35.7% was expected because predictive values are not intrinsic to the test but rather depend on the prevalence of the disease screened for, with PPV directly proportional to prevalence [67]. Fredrick et al. [61] in their study of a similar population documented a higher sensitivity, lower specificity, higher PPV, and slightly lower NPV of 76.6%, 85.9%, 52.1% and 94.8% respectively, compared with LE test performance in this study. The lower sensitivity in the index study could be due to a number of factors associated with false negative test results, such as high specific gravity of urine, ascorbic acid, ketonuria, and proteinuria, which were not considered in the interpretation of LE results in the index study. The lower PPV is likely due to a much lower prevalence of UTI in this study compared with that in the study by Fredrick et al. [61] (4.8% vs. 16.8%).

The nitrite test had a low sensitivity of 25% and a specificity of 100% in this study, meaning a positive nitrite test in the study population is strongly indicative of UTI and justifies commencement of empirical treatment. This is very similar to reports by Rabasa et al. [8] and Festo et al. [68] in the corresponding study populations in Maiduguri and Tanzania, respectively. The low sensitivity in this study is as expected in infants and young children because production of nitrite in the urinary bladder requires approximately four hours, which is not easily obtainable in this age group, particularly infants, who void frequently [21]. Therefore, a negative test should be interpreted with care and has little value in excluding UTI. Furthermore, isolation of Staph. aureus, which is a non-nitrate-reducing bacterium, contributed to the low sensitivity. The PPV and NPV of the nitrite test were 100% and 96.3% respectively, but these predictive values should be interpreted with caution because of the effect of the prevalence of UTI in this study.

The sensitivity and specificity of dipstick urinalysis (a combination of LE and/or nitrite) in this study were 37.5% and 97.1% respectively. This increased the sensitivity of both LE (31.3%) and nitrite (25%), whereas the specificity of nitrite (100%) was slightly reduced, and that of LE (97.1%) was unaffected. The observed effect of the combination of both tests on their sensitivity and specificity is consistent with reports from earlier studies [65, 68, 69]. Therefore, a negative dipstick result is superior in ruling out disease than a negative test of the individual components, though a sensitivity of 37.5% still implies that dipstick urinalysis is not good for ruling out UTI in this study population.

Microscopy for pyuria (uncentrifuged urine) showed better sensitivity (62.5%) than LE strips because visualisation of WBCs when present in urine is influenced to a lesser degree than identification of WBC using LE test strips. Also, the LE test strip had a practical detection limit of 10 to 25 WBCs/microlitre of urine, whereas microscopy identifies pyuria at lower counts.

Gram stain of uncentrifuged urine had a high sensitivity, specificity, PPV, and NPV of 93.8%, 99.7%, 93.8% and 99.7% respectively. The high sensitivity and specificity are likely because all the uropathogens responsible for UTI in this study were gram-detectable, and sample collection was done carefully to avoid contamination and consequent increase in false positive results. Musa-Aisien et al. [9] reported a sensitivity, specificity, PPV, and NPV of 77%, 98%, 77% and 98% respectively, in a related study. The difference in sensitivity compared to the index study may be due to the different competence levels of the lab collaborators engaged. Microscopy of Gram-stained urine has been reported to be the most sensitive single test on urinalysis for the diagnosis of UTI in children [69]. The high sensitivity and specificity recorded in this study make it a very good test for ruling in and ruling out UTI in febrile children, although the expertise required in its performance may make it difficult for some facilities, mainly in resource-poor settings, to routinely carry it out.

Combined uncentrifuged microscopy (pyuria + gram stain) also had high accuracy (93.8% sensitivity; 95.5% specificity). Urine microscopy of uncentrifuged urine, when possible, should be considered in the review of febrile under-fives with fever, and it becomes more valuable when the dipstick test is negative. Centrifugation of urine before microscopic examination for pyuria yielded a sensitivity, specificity, PPV, and NPV of 56.3%, 96.8%, 47.4% and 97.8% respectively, in this study, which is about the same accuracy as microscopy for uncentrifuged pyuria noted earlier. On this basis, the extra steps involved in centrifugation with the attendant time spent are unnecessary, making it possible for microscopy results to be available earlier. Festo et al. [68] documented a sensitivity and specificity of 44.2% and 86.5% which is lower than the results of this study, probably due to the expertise involved in the test performance in the index study.

Gram stain of uncentrifuged urine had the best accuracy of the single tests that can be performed on urinalysis for UTI, while microscopy of uncentrifuged urine also had good accuracy. If these tests are used effectively, they will help reduce the number of children with UTI left untreated. It will also decrease the number of unnecessary cultures ordered as well as unnecessary empirical treatments.

The commonest uropathogen in this study was Escherichia coli, which was isolated in 43.8% of subjects with UTI, consistent with findings from studies of similar populations in Nigeria, [7, 9] Tanzania [68] and India [50, 51]. This is followed by Klebsiella spp. (37.5%) in line with the findings by Musa-Aisien et al. [9] in Benin and Fredrick et al. [61] in Tanzania, but in contrast to the study by Ibeneme et al. [7] in Enugu, where Staph. aureus followed E. coli as the commonest organism. UTI is commonly a result of ascending infection caused by faecal flora, principally gram-negative colonic bacteria, as seen in this study.

Antibiotic susceptibility pattern of the isolated organisms showed that they were all sensitive to Meropenem, an ultra-broad-spectrum antibiotic which is relatively expensive and therefore not commonly abused. Cefepime and Ceftazidime also had high in-vitro activity against most of the organisms and should be the preferred empirical treatment for febrile children with suspected UTI based on clinical assessment and/or urinalysis while waiting for culture and sensitivity results to guide management. The very poor sensitivity of the organisms to cefuroxime and ampicillin is likely due to the high rate of abuse with consequent development of resistance. Understandably, there are wide differences in the reported sensitivity patterns of organisms responsible for UTI, which is why it is recommended that empirical treatment should be dependent on local susceptibility studies like the index study [21].

Conclusion

UTI is common in febrile children under five presenting to NAUTH Nnewi, especially females, with higher prevalence in infant males. Many cases are missed due to non-specific symptoms, often mistaken for malaria. An axillary temperature ≥ 38.4 °C is significantly associated with UTI and makes it a key clinical indicator. While dipstick urinalysis for leukocyte esterase and nitrite offers high specificity, its sensitivity is low, making it useful for ruling out individuals without UTI. Microscopy of uncentrifuged urine (pyuria and Gram stain) performed better than dipstick, with Gram stain having the highest accuracy among the individual tests, aiding diagnosis, especially when dipstick results are negative. Urine culture remains the definitive diagnostic test. E. coli is the most common pathogen, with antibiotics like Meropenem, Cefepime, and Ceftazidime showing good activity for empirical treatment. In line with these findings, doctors attending to febrile under-fives should continually be sensitised to the possibility of underlying UTI in these patients to reduce the level of missed diagnosis. Dipstick urinalysis, which is affordable, easy to perform, and gives immediate results, should be adopted as a routine test in febrile under-fives.

Limitations of the study

1. The study focused on children with documented fever on presentation, which may have led to the wrongful exclusion of children with UTI and those with a history of fever that were probably given antipyretics before presentation.

2. Investigations to rule out other causes of fever in the subjects with UTI were not included in this work due to limited funds.

3. A limitation of the study is the low number of children with confirmed UTI, which reduces the clinical utility of the results; however, the observed high specificity of the dip stick and microscopy tests will minimise false-positive results, and in turn add the needed value in such a low-prevalence setting.

4. Sample collection as expected in children was challenging, and about one-third of the samples were collected via urethral catheterisation, which is inferior to mid-stream and supra-pubic aspiration methods of urine collection.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AAP

American Academy of Paediatrics

AIDS

Acquired Immunodeficiency Syndrome

CFU

Colony Forming Unit

CHER

Children Emergency Room

CHOP

Children Outpatient

CKD

Chronic Kidney Disease

CNS

Central Nervous System

CSU

Catheter Specimen Urine

CVU

Clean Voided Urine

DM

Diabetes Mellitus

ESRD

End Stage Renal Disease

HIV

Human Immunodeficiency Virus

HPF

High Power Field

LE

Leukocyte Esterase

NAUTH

Nnamdi Azikiwe University Teaching Hospital

OIF

Oil Immersion Field

RTI

Respiratory Tract Infection

SEC

Socioeconomic Class

SG

Specific Gravity

SN

Sensitivity

SP

Specificity

SPA

Suprapubic Aspiration

UTI

Urinary Tract Infection

VUR

Vesicoureteric Reflux

WBC

White Blood Cell

WHO

World Health Organization

PPV

Positive Predictive Value

NPV

Negative Predictive Value

Author contributions

Conception – CCN, IE, JCE Design of the work – CCN, IE, JCE, INA Acquisition and analysis –CCN, INA, CON Interpretation of data - CCN, IE, JCE, CNA, INA CON Drafting/Revision of the paper – CCN, CON. All authors read and approved the submitted version and agree to be responsible and accountable for their contribution to the work.

Funding

No external funding was received for this study.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethical approval

Ethical approval was obtained from the Research Ethics Committee of the Nnamdi Azikiwe University Teaching Hospital, Nnewi.

Consent for publication

Informed consent was obtained from parents in compliance with the Declaration of Helsinki.

Competing interests

The authors declare no competing interests.