Hyperbilirubinemia and urinary tract infection: the effect of indirect hyperbilirubinemia on the in vitro growth of uropathogen Escherichia coli in newborn urine

Abstract

High serum bilirubin is antioxidant and cytoprotective. We evaluated if urine samples of hyperbilirubinemic newborns impede uropathogenic Escherichia coli growth. Bag-urine samples of hyperbilirubinemic newborns (study group) were cultured at presentation and during remission. Urine sample were obtained only once from healthy newborns (control group). Escherichia coli [2 × 104 colony-forming unit (cfu)/mL] was inoculated into the sterile urine samples and colony counts were determined after 24 h. Bilirubin levels at presentation and remission were also recorded. Escherichia coli colony counts of the control versus study groups and of the presentation versus remission samples in the study group were compared. There were 13 study and 17 control cases. Escherichia coli colony counts were not different in the study group at presentation versus remission (5.4 ± 0.7 vs. 5.5 ± 0.8 log₁₀, respectively; p = 0.659). Escherichia coli colony count of the control group (5.2 ± 0.6 log₁₀) was also not different from the study group. In conclusion, the urine of hyperbilirubinemic newborns did not affect the growth rate of uropathogenic E. coli.

• Escherichia coli
• hyperbilirubinema
• jaundice
• newborn
• urinary tract infection

Introduction

Hyperbilirubinemia is a frequent neonatal problem. Physiologic jaundice in the first week of life is present in over half of newborns and reflects the transition from an intrauterine to an extrauterine pattern of bilirubin transport and metabolism. Serum bilirubin levels typically peak between 3 and 5 days of age and usually improve by 7 days of age. However, in breastfed infants, jaundice can persist beyond 10–14 ays of age.1

Urinary tract infections (UTIs) have been documented in previous studies as a cause of neonatal jaundice.2–5 Recently, asymptomatic, jaundiced infants <8 weeks old were reported to have a rate of UTI that is as high as the rate in febrile infants of same age.6 Hyperbilirubinemia has generally been regarded as a pathologic consequence of UTIs, but no clinical study has been conducted to evaluate the effects of hyperbilirubinemia on the course of UTIs.

We hypothesized that urine of hyperbilirubinemic newborns might include some factors that impede growth of these bacteria in newborns vulnerable to UTI and its sequelae. Thus, in this study we aimed to evaluate the effect of the urine of hyperbilirubinemic newborns on the growth potential of an uropathogenic E. coli strain.

Patients and methods

Selection of patients and control subjects

The data of term newborn babies presenting with indirect hyperbilirubinemia over 15 mg/dL within the first week of life admitted to Dr. Ekrem Hayri Üstündağ Obstetrics and Gynecology Hospital and to Izmir Tepecik Training and Research Hospital for evaluation of any underlying etiology and for phototherapy were collected. Detailed prenatal, natal and postnatal history of the patients were obtained regarding infectious or systemic diseases of the mothers during pregnancy, gestational age with respect to last menstrual date, birth weight, APGAR scores, postnatal age, onset of jaundice and type of feeding (breast milk or formula).

A bilirubin work-up including complete blood count, peripheral smear, reticulocyte count, direct Coombs test, venous (total and direct) and capillary (total) bilirubin levels, urinalysis and glucose-6-phosphate dehydrogenase (G-6PD) level were performed in all patients. Capillary blood bilirubin levels during remission period (after 2–3 days of phototherapy) were also recorded. Breastfed newborns without any underlying pathology other than hyperbilirubinemia were considered eligible for the study group. Only breastfed babies were selected in order to standardize the patients as human milk has been reported to provide a multifactorial support system for the babies including protective factors against UTIs,7 although not supported by a recent study.8 A control group of newborn babies without hyperbilirubinemia was formed. These babies were selected among the unjaundiced breastfed babies presenting for the regular 15th day visit to the same hospital.

Urine cultures and inoculation of E. coli

Two bag-urine samples were obtained for culture from the study patients, one at presentation to the hospital (at 2–4 days of age, when serum bilirubin was higher), and the other during remission (after 2–3 days of phototherapy). Only one urine sample was obtained for culture from the control subjects. All urine samples were transported within 30 min in ice bags to the Dokuz Eylül University Medical Faculty Microbiology Laboratory. All samples were handled by the same microbiologist who was unaware of the patient group (study or control) and timing of culture (at presentation or during remission). Urine samples were divided into two sterile tubes (5 mL each) and one was cultured on a standard sheep blood agar without any further processing. A uropathogenic E. coli strain (ATCC 25922) was inoculated to the other sample. Inoculum was prepared from a 24-h culture as a 0.5 McFarland standard bacterial suspension. Serial dilutions were made to determine the number of microorganisms and 0.1 mL of each dilutions were cultured on blood agar plates. After that, 0.1 mL of suspension including 2 × 10⁵ colony-forming unit (cfu)/mL E. coli was added to 0.9 mL of urine sample and incubated for 4 h. Then, 0.1 mL of this suspension was inoculated on blood and EMB agar plates. After incubation of 24 h at 37 °C, colony numbers were determined as cfu/mL.9 Urine samples were discarded if there was any growth of bacteria in the unprocessed urine (the sample into which E. coli was not inoculated).

The study was approved by the Ethical Committee of Dokuz Eylül University Medical Faculty. Written consents of the parents were obtained before participation in the study.

Statistical evaluation

Serum bilirubin levels and E. coli cfu counts at presentation and during remission in the study group were compared by Wilcoxon signed rank test. Serum bilirubin levels and E. coli cfu counts of the study and control groups were compared by Mann–Whitney U-test. Correlation of capillary blood bilirubin levels and E. coli colony counts in the study group was made by Spearman’s test. A p-value <0.05 was considered as significant.

Results

A total of 23 full-term, breastfed, jaundiced newborns without any identifiable cause of hyperbilirubinemia were enrolled in the study group. There were 29 participants in the control group. However, 10 patients in the study group and 12 cases in the control group were excluded due to bacterial contamination of their bag-urine samples (all bacterial growth in these cases were either <10⁵ cfu/mL of a single organism or growth of more than one organism). Study and control groups were not different with respect to gender (male/female = 7/6 vs. 10/7, respectively; p = 0.785) and birth weight (3371 ± 429 vs. 3212 ± 318, respectively; p = 0.300). The bilirubin work-up of patients did not demonstrate any significant underlying disorder. None of the infants had a high direct bilirubin level. Urinalysis was normal in all urine samples of the study and control cases.

Mean total bilirubin levels of the study group and mean E. coli colony counts of both groups were presented in Table 1. Total bilirubin levels decreased significantly during remission in the study group. However, growth of E. coli did differ neither within the study group (at presentation vs. during remission) nor between the control and study groups. Serum bilirubin levels and E. coli cfu counts (at presentation and during remission) were not correlated in the study group (n = 20, r = −0.019, p = 0.936).

Table 1. Capillary blood total bilirubin levels in study group and E. coli colony counts in study and control groups.

	Study group
	At presentation (n = 13)	During remission^a (n = 7)	Control group (n = 17)
Total bilirubin^b (mg/dL)	17.5 ± 1.5	8.9 ± 0.8	–
E. coli growth^c (cfu/mL log10)	5.4 ± 0.7	5.5 ± 0.8	5.2 ± 0.6

^aSix of the urine samples obtained during remission were contaminated. Thus, only seven patients were evaluated during remission period.

^bp = 0.018.

^cThere was no difference within study group (at presentation versus during remission, p = 0.893). Control group was also not different from the study group (p = 0.659 and p = 0.708 at presentation and during remission, respectively).

Discussion

High bilirubin levels, a prevalent condition among newborns, have antioxidant and cytoprotective properties.10–12 As newborns have high rate of bacterial infections, hyperbilirubinemia may have beneficial effects in this age group. Newborns also have high rate of UTIs and of renal scarring following pyelonephritis.13,14 Garcia and Nager6 reported that infants with the reported onset of jaundice after 8 days of age, when physiologic jaundice is expected to have improved or resolved, had a higher incidence of UTIs. Moreover, it has been stated that newborns over 3-day of age with indirect hyperbilirubinemia (total bilirubin >15 mg/dL) have a UTI prevalence of 8%, while none of them had renal scarring.15 Thus, it seems possible that hyperbilirubinemia may develop as a protective mechanism in these babies against pyelonephritis and renal scarring. On the other hand, although urine is an adequate medium for growth of uropathogenic E. coli strains,16 several factors required for bacterial growth are limited in urine and must be provided by de novo bacterial synthesis. It has been proposed that agents specifically inhibiting synthesis of one or more of these factors may have therapeutic value for prevention or treatment of UTIs.17

Hyperbilirubinemia in UTIs has been reported to be unconjugated and related to hemolysis caused by E. coli and other Gram-negative organisms, or conjugated secondary to cholestasis.4,5,18 However, the relationship between UTI and hyperbilirubinemia has not been clearly defined.19 We proposed that hyperbilirubinemia might have protective effects in UTI (due to antioxidant and cytoprotective effects or regarding inhibition of the growth of urinary pathogens). However, the results of this preliminary study did not show any difference in growth rate of uropathogenic E. coli in the urines of hyperbilirubinemic newborns at presentation and during remission. In addition, newborns without jaundice were not different from the hyperbilirubinemic newborns with respect to E. coli growth rate in their urine. We did not enroll newborns with bacterial growth in their unprocessed urine. However, as none of these babies had real UTI (all were contamination), elimination of these babies does not indicate that patients having jaundice associated with UTI were excluded from the study creating a selection bias.

Although in vitro uropathogenic E. coli growth seems to be unaffected by the presence or absence of hyperbilirubinemia or by the degree of hyperbilirubinemia, in vivo effects of hyperbilirubinemia on UTIs may still be present. Hyperbilirubinemia in newborn babies has been assumed to have protective role for renal scar formation due to UTI as these babies has high rate of UTI and their immature kidneys are more vulnerable to infectious damage. In accordance with this assumption, exogenous bilirubin administration to the rats with experimentally induced pyelonephritis prevented renal scarring significantly when used in combination with antibiotics.20 On the other hand, Xinias et al.21 stated that hyperbilirubinemia is an indicator of renal damage in newborns with UTI. Furthermore, it has been stated that high bilirubin levels cause relaxation of ureter and bladder musculature in vitro and thus hyperbilirubinemia might predispose to UTI development due to difficulty in urinary passage, urine retention and stasis in jaundiced infants.22 Thus, further studies seem to be needed to better clarify the association of hyperbilirubinemia and UTIs in newborns.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

Acknowledgements

We appreciate the contributions of Esra Özer, MD from the Izmir Tepecik Training and Research Hospital, Turkey.