Objective. Urinary tract infections (UTIs) are a common clinical problem in febrile infants younger than 8 weeks old, with a prevalence between 5% and 11%. Previous studies have noted that jaundice may be one of the first signs of a bacterial infection in infants. Our goal was to determine the incidence of UTIs in asymptomatic, jaundiced infants younger than 8 weeks old and to determine which historical and laboratory parameters are associated with UTIs.
Methods. We prospectively evaluated asymptomatic, jaundiced infants younger than 8 weeks old for evidence of a UTI (defined as >10 000 colony-forming units per milliliter of a single pathogen, obtained by bladder catheterization). A serum fractionated bilirubin level was obtained on all study patients. Detailed questionnaires were completed, which included demographic information, prenatal, intrapartum, and postnatal events.
Results. Twelve (7.5%) of 160 infants had a UTI (95% confidence interval: 3.9%–12.7%). Isolated organisms included Escherichia coli, Enterobacter cloacae, Enterococcus, Klebsiella pneumoniae, group B Streptococcus, Streptococcus viridans, and Staphylococcus aureus. Patients with the onset of jaundice after 8 days of age had a higher incidence of UTI, 6 (50%) of 12 infants in the positive culture group, versus 15 (10%) of 148 infants in the negative culture group. Abnormal urinalysis and microscopy results were noted in 5 (42%) of 12 infants with a UTI, and 6 (55%) of 11 infants had abnormal renal ultrasound results.
Conclusion. A UTI was found in 7.5% of asymptomatic, afebrile, jaundiced infants younger than 8 weeks old. In addition, infants with the onset of jaundice after 8 days of age or patients with an elevated conjugated bilirubin fraction were more likely to have a UTI. Therefore, we recommend that testing for a UTI be included as part of the evaluation in asymptomatic, jaundiced infants presenting to the emergency department.
Each year 60% of newborns become clinically jaundiced.1 The American Academy of Pediatrics (AAP) has published guidelines outlining the management of healthy newborns with hyperbilirubinemia, which includes maternal ABO and Rh typing, direct Coombs’ test, blood and Rh (D) typing of infants’ cord blood, and a total serum bilirubin level.1 If jaundice persists beyond 3 weeks of age in healthy infants, measurement of total and direct serum bilirubin is advocated.1 Although some reports have suggested that unexplained hyperbilirubinemia may be associated with bacterial infections in newborn infants, the AAP guidelines do not recommend an evaluation for such infections.
Urinary tract infections (UTI) occur with a prevalence between 5% and 11% among febrile infants younger than 8 weeks old.2–5 The incidence of bacteremia and sepsis in an infant with a UTI has been reported to be between 6% and 36%.2,4–6 The incidence of urinary tract abnormalities in infants with UTIs is between 30% and 55%, the most common being vesicoureteric reflux.6–11 The majority of UTIs in young infants at the time of diagnosis are thought to represent acute pyelonephritis.10 The clinical manifestations of UTIs in newborn infants are extremely variable, ranging from severe illness to nonspecific signs and symptoms, such as growth failure, vomiting, diarrhea, fever, irritability, lethargy, abnormal urination with oliguria, polyuria or malodorous urine, and jaundice.12
The goals of this study were 1) to determine if unexplained hyperbilirubinemia is associated with a UTI in infants; 2) to determine the incidence of a UTI in asymptomatic jaundiced infants; and 3) to determine which historical and laboratory parameters are associated with a UTI in asymptomatic, jaundiced infants younger than 8 weeks old.
Jaundiced infants younger than 8 weeks old, who presented to the emergency department (ED) at the Childrens Hospital Los Angeles from February 1998 through February 2000 were eligible for study. Clinical jaundice was defined as any yellow or green-yellow discoloration of the skin, mucous membranes, or sclera. Patients were excluded if they had fever by history, documented fever >38°C, vomiting, diarrhea, poor feeding, tachypnea, lethargy, or irritability. Patients were enrolled 24 hours a day, 7 days a week. The study was reviewed and approved by the Committee on Clinical Investigation (Institutional Review Board) of Childrens Hospital Los Angeles, and written informed consent was obtained from the caregivers of all study participants.
Detailed questionnaires were completed by the resident or attending physician on-duty in the ED. Demographic information was collected, including prenatal, and intrapartum events, such as gestational age, maternal infections, pregnancy complications, prolonged rupture of membranes, mode of delivery, and maternal fever. Postnatal events including neonatal fever, previous evaluation for neonatal sepsis, and age at discharge from the newborn nursery, were also recorded. Other historical factors included the onset of jaundice and progression of jaundice as reported by parents, and whether the infant was breastfed versus formula-fed.
All study participants had a urine sample obtained by bladder catheterization. Part of the urine sample was tested in the ED using Multistix 10 (Bayer, Elkhart, IN). The remainder was tested in the clinical laboratory by urinalysis, microscopy, Gram stain, and culture. In the laboratory, a standard urinalysis was performed using Multistix 10 (Bayer), and urine specimens were centrifuged at 2000 revolutions per minute for 5 minutes, resuspended and stained; examined microscopically under high-power field (HPF) for pyuria, reported as leukocytes per HPF and bacteriuria, reported as none, few, moderate, and many. All urine specimens were sent for standard quantitative culture and considered positive if a single pathogen ≥10 000 colony-forming units was isolated.13 In addition, a serum fractionated bilirubin level was obtained on all study participants.
Disposition and Follow-up
After an initial evaluation in the ED, patients were either discharged from the hospital to follow-up with their primary care physician or were admitted to the hospital for phototherapy and/or a sepsis evaluation at the discretion of the ED attending physician. If at a later date, the urine culture became positive, patients were contacted and instructed to return to the ED for a sepsis evaluation and admission to the hospital. The sepsis evaluation included a complete blood cell count, standard urinalysis with microscopy, lumbar puncture, and cultures of blood, urine and cerebrospinal fluid. After diagnosis of a UTI, the infant’s primary care physician or the general pediatric service resident and attending physician managed the patient while in the hospital. Follow-up information was obtained from medical record review.
Sample-size estimates were based on an incidence of bacteriuria that ranges from 0.5% to 1% in asymptomatic, term infants.14–17 To detect a difference of 5%, with 95% confidence interval (CI) and a power of 80%, a sample size of 90 patients or more was needed.18
Questionnaires were created using Teleform 6.1,19 after which data were key-entered and verified using Teleform Verifier.19 The data were then validated and analyzed using SAS 6.12 statistical software.20 Descriptive data were reported as mean ± standard deviation and associations between categoric variables were analyzed using the Fisher exact test or the χ2 test, and continuous variables using Student t test. Statistical significance was defined as a P value <.05. CIs were also calculated.
Study Population and Demographics
From February 1998 through February 2000, 187 asymptomatic, jaundiced infants were cared for in the ED at Childrens Hospital Los Angeles, and were eligible for recruitment. The parents of 14 infants (7%) declined to participate, and the ED attending physician on-duty did not offer study participation to the caregivers of 11 infants (5%). One patient was withdrawn because of improper urine collection and 1 patient was withdrawn secondary to maternal syphilis. Therefore, our study population consisted of 160 infants. The mean age of the study population was 12.1 days ± 9.6 days of age. The mean birth weight was 3.27 kg ± 0.58 kg. Ninety percent of infants were born at term (37–42 weeks’ gestational age), and 85% were born via vaginal delivery. Ninety-eight infants (61%) were males and 62 infants (39%) were females. Latinos comprised 79% of study participants, which is comparable to the ethnic distribution of patients seen in our ED. The remaining participants were Asian/Pacific Islanders (10%), African American (4.5%), white (4%), and other (2.5%). Demographic and historical characteristics of patients with positive and negative urine cultures are described in Table 1. No statistically significant differences were found between groups with regard to age, gender, birth weight, gestational age, mode of delivery, or maternal complications, such as fever, premature labor, prolonged rupture of membranes, or previous infection. Four males were circumcised and none had a UTI. Although not statistically significant, positive urine cultures were less frequent in breastfed infants as compared with formula-fed infants.
Incidence of UTIs
Of 160 participants enrolled, positive urine cultures were found in 12 infants (7.5%; 95% CI: 3.9%–12.7%). Bacterial pathogens isolated from urine cultures are presented in Table 2. Single pathogens were isolated from the initial urine culture in 8 of 12 study participants. The remaining 4 participants had an initial urine culture which grew ≥2 organisms, however, urine culture testing on reevaluation and before starting antibiotics demonstrated a single organism in each of 4 patients (Escherichia coli [2 patients], Enterobacter cloacae, and group B streptococcus). E coli was isolated from the urine in 4 of 12 infants with a positive urine culture, Enterococcus from 2 patients, E cloacae from 2 patients, and Klebsiella pneumoniae, group B Streptococcus, Streptococcus viridans, and Staphylococcus aureus from 1 patient each.
Urinalysis and Microscopy
Standard urinalysis, microscopy, and Gram stain results in infants with positive urine cultures are reported in Table 2. Urinalysis results were positive for leukocyte esterase and blood in 5 participants (42%), one of which was positive for nitrites. Centrifuged urine specimens examined microscopically under HPF were positive for bacteria in 5 (42%) of 12 infants. Of those 5 patients, 2 infants (17%) had both bacteriuria and pyuria defined as ≥10 white blood cells (WBC) per HPF (WBC/HPF).13 Of note, 7 infants (58%) with a positive urine culture had a negative urinalysis or ≤5 WBC/HPF on urine microscopy.
Degree of Jaundice
Jaundice before postpartum hospital discharge was reported by the parents in 69 (43%) of 160 infants. Table 3 shows the degree of jaundice and bilirubin levels in patients with positive and negative urine cultures. Sixty percent of infants had jaundice reported before 3 days of age. Patients with the reported onset of jaundice after 8 days of age had a higher incidence of UTIs, 6 (50%) of 12 infants in the positive culture group, versus 15 (10%) of 148 infants in the negative culture group (P < .001). Although not statistically significant, a history of worsening or progressive jaundice was observed more commonly in the positive culture group, 8 (67%) of 12 infants, versus 65 (44%) of 148 infants in the negative culture group. Total bilirubin levels and conjugated bilirubin fraction are presented in Table 2. There were no differences in mean total bilirubin levels in breastfed infants (13.9 mg/dL), as compared with formula-fed infants (13.1 mg/dL). Total bilirubin levels were not higher in patients with UTIs as compared with patients with negative urine cultures. An increase in the conjugated bilirubin fraction was found in 2 (17%) of 12 patients with a positive urine culture, while none were noted in the negative culture group (P < .01). The 2 patients with an increase in the conjugated bilirubin fraction were a 32 day-old female with an E coli UTI with a conjugated bilirubin fraction of 12.5 mg/dL, and a 44-day-old male with K pneumoniae UTI with a conjugated bilirubin fraction of 3 mg/dL.
After the initial study evaluation was completed, 25 patients were admitted to the hospital. Twenty-two patients were admitted secondary to jaundice (total bilirubin level ≥20 mg/dL), and 3 patients were admitted with a presumed UTI, as suggested from the urinalysis and microscopy results. Bacterial pathogens in the urine were isolated from 5 (20%) of 25 patients initially admitted to the hospital. Eight infants were reevaluated in the ED after urine cultures became positive, and all were subsequently admitted to the hospital for a sepsis evaluation and treatment. One patient, a 32-day-old female had a UTI and bacteremia, both secondary to E coli. Cerebrospinal fluid cultures were negative on all patients with positive urine cultures.
Although not part of the study design, 11 of 12 infants diagnosed with a UTI had renal ultrasounds performed. One patient did not return for imaging studies. Urinary tract abnormalities were found in 6 (55%) of 11 patients, which included hydronephrosis (3 patients), pelviectasis (2 patients), and renal stone (1 patient). No abnormalities were found in voiding cystourethrograms obtained in 4 infants.
Bacteremia and sepsis have been well-documented in previous studies as a cause of neonatal jaundice in seriously ill newborns. Multiple studies have described patients with proven bacterial infection, who developed jaundice during the course of their illness.21–24 Other studies, have noted that jaundice may be one of the first signs of bacterial sepsis in neonates in the first few days of life. Linder et al,25 in a prospective study of 5805 full-term newborns, identified 3 of 93 infants younger than 7 days old, who had unexplained jaundice and sepsis. Mora and Rodriguez26 reported similar findings in 16 patients with bacteremia, in which jaundice was the primary presenting sign.
Other studies have indicated that jaundice may be one of the first signs of a UTI. In Seeler’s expanded series27–29 in 1977, 24 bottle-fed infants were discharged from the newborn nursery, but returned with moderate signs and symptoms of illness, and later found to have a UTI. Chavalitdhamrong et al,30 in a small series of 69 asymptomatic neonates with unexplained jaundice, found evidence of Gram-negative UTIs in 2 infants. Rooney et al, 31 reported results on a series of 22 jaundiced infants with documented bacterial infections, with mild to moderate signs or symptoms of illness. A UTI was present in 9 infants (40%). Maisels and Kring32 in 1992 performed a retrospective study and found no cases of sepsis in a series of 306 patients, who were admitted to the hospital for evaluation of indirect hyperbilirubinemia during the first week of life. However, only 126 infants (41%) had the urine tested for infection, thus possibly missing cases of UTIs. In addition, the mean age of that study population was 5 days of age, in contrast to our study, in which the mean age was 12.1 days. To our knowledge, no previous study has looked at the incidence of UTI in asymptomatic jaundiced infants younger than 8 weeks old. In the present study, a UTI was found in 7.5% of asymptomatic, jaundiced infants younger than 8 weeks old. Interestingly, similar results are reported in febrile infants younger than 8 weeks old, with a prevalence of a UTI between 5% and 11%.2–5 In our study, the most common pathogen isolated was E coli, followed by E cloacae and Enterococcus species.
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 to 5 days of age and usually improves by 7 days of age. However, in breastfed infants, jaundice can persist beyond 10 to 14 days of age.33 In our study, >60% of infants had the onset of jaundice before 3 days of age. However, 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. Although a history of worsening jaundice as reported by parents was more commonly observed in patients with positive urine cultures, it was not statistically significant. In addition, total serum bilirubin levels were similar in patients with a UTI, as compared with those infants with negative urine cultures. However, an increase in the conjugated bilirubin fraction was noted in 2 of 12 infants with a positive urine culture, while none of the 148 patients with a negative urine culture had an elevation in the conjugated bilirubin fraction. After treatment of the UTI, both infants with an increase in the conjugated bilirubin fraction had resolution of their jaundice requiring no additional work-up. The hyperbilirubinemia associated with UTIs can be unconjugated and related to hemolysis caused by E coli28,34 and other Gram-negative organisms,29 or conjugated secondary to cholestasis.21,22 The mechanism by which a UTI causes cholestasis is not clear, but possible mechanisms include microcirculatory changes in the liver, direct effects from bacterial products, and/or from endotoxin-induced mediators.35,36 It is postulated that even mild hemolysis can overload the immature liver conjugating mechanism, leading to an increase in serum bilirubin levels.
Is it possible that patients could have asymptomatic bacteriuria rather than a UTI? Asymptomatic bacteriuria is a condition in which the urogenital tract is thought to be colonized rather than infected. It is seen in young white females, but uncommon in young males. Previous studies by Abbott et al,14 Edelman et al,15 Drew et al,16 and Wettergren et al17 have found the incidence of bacteriuria in asymptomatic neonates to be between 0.5 and 1%.17–20 Pyuria, the presence of WBC in the urine has traditionally been the marker for distinguishing a true UTI from asymptomatic bacteriuria.37 However, pyuria is not a sensitive marker in neonates. Crain and Gershel,3 in a prospective study of 430 febrile infants younger than 8 weeks old, identified 33 with a UTI. Only 48% of infants had an abnormal urinalysis, defined as 5 or more WBC/HPF or visible bacteria on microscopic examination of unspun urine.3 Similar studies by Landau et al38 reported that 31 (28%) of 128 infants younger than 16 weeks old with a UTI, had fewer than 5 WBC/HPF in the urine. Dimercaptosuccinic acid renal scintigraphy radionuclide scanning indicated pyelonephritis in 4 (13%) of 31 infants without pyuria.38 We report similar results in our study, in which 50% of asymptomatic, jaundiced infants with a UTI, had an abnormal urinalysis or microscopy results (defined as 10 or more WBC/HPF, positive leukocyte esterase, and/or a positive Gram stain).
In a recent prospective study of infants younger than 8 weeks old diagnosed with a UTI, 22 (49%) of 45 infants had urinary tract abnormalities that included hydronephrosis, renal atrophy, and vesicoureteral reflux.39 Other studies have also reported the prevalence of urinary tract abnormalities between 30% and 55%.6–11 Although not specifically studied, we found urinary tract abnormalities in 6 (55%) of 11 infants evaluated by renal ultrasound, which included hydronephrosis, pelviectasis, and renal stone.
Our observation, that the incidence of a UTI in asymptomatic, jaundiced infants was approximately the same as febrile infants younger than 8 weeks old, suggests that we may be identifying infants with a UTI before signs and symptoms become evident. We do not consider our positive urine cultures to reflect incidental asymptomatic bacteriuria for several reasons. Firstly, our incidence of 7.5% was substantially higher than the reported 0.5% to 1% by other studies in symptom-free infants.17–20 Secondly, the incidence of urinary tract abnormalities that we detected is similar to that described in previous studies, supporting the fact that infants with urinary tract abnormalities are predisposed to, a UTI. Finally, the true test of asymptomatic bacteriuria would be to withhold antibiotics and observe for clinical signs and symptoms of infection. Ethically, not treating a “positive urine culture” in infants younger than 8 weeks old, to test this hypothesis was not felt to be reasonable, or risk-free.
Our study had several limitations including missed study participants that were not offered study participation. In addition, not all infants with positive urine cultures had a renal ultrasound and voiding cystourethrogram, and none had a dimercaptosuccinic acid renal scintigraphy radionuclide scan. Also, since the majority of infants were not circumcised in our study, we can not draw conclusions regarding circumcision and the incidence of UTIs. This, however, may be a significant factor that should be investigated in future related studies. Given the fact that our study was conducted solely in the ED, the results obtained can not be generalized to patients in other outpatient settings.
UTIs can occur in asymptomatic, jaundiced infants. In this study, the incidence of a UTI in asymptomatic, afebrile, jaundiced infants younger than 8 weeks old was 7.5%. Infants with the onset of jaundice after 8 days of age, when physiologic jaundice is expected to have improved or resolved, were more likely to have a UTI. In addition, although our numbers are small, all patients with an increase in the conjugated bilirubin fraction were diagnosed with a UTI. Finally, our results show that the incidence of a UTI in asymptomatic, afebrile, jaundiced infants is similar to the incidence of UTI in febrile infants. Thus, jaundice may be the first sign of a UTI in asymptomatic infants before other signs and symptoms become evident. Therefore, we suggest that testing for a UTI be part of the diagnostic evaluation of asymptomatic, jaundiced infants presenting to the ED.
This study was supported in part by the National Institute of Health National Center of Research Resources General Clinical Research Center grant MO1RR-43 and was performed in part at the General Clinical Research Center at the Childrens Hospital Los Angeles.
We thank Earl Leonard for his biostatistical support and Beth Haden, RN, for her assistance. We thank Bard Medical for donating the urinary catheters used on study participants. We also thank the housestaff at Childrens Hospital Los Angeles and the nursing and ancillary staffs of the Emergency Department for their invaluable assistance.
- ↵American Academy of Pediatrics, Provisional Committee for Quality Improvement and Subcommittee on Hyperbilirubinemia. Practice parameter: management of hyperbilirubinemia in the healthy term newborn. Pediatrics.1994;94 :558– 565
- ↵Krober MS, Bass JW, Powell JM, Smith FR, Seto DS. Bacterial and viral pathogens causing fever in infants less than 3 months old. Am J Dis Child.1985;139 :889– 892
- ↵Crain EF, Gershel JC. Urinary tract infections in febrile infants younger than 8 weeks of age. Pediatrics.1990;86 :363– 367
- ↵Ginsburg CM, McCracken GH. Urinary tract infections in young infants. Pediatrics.1982;69 :409– 412
- Bourchier D, Abbott GD, Maling TMJ. Radiological abnormalities in infants with urinary tract infections. Arch Dis Child.1984;59 :620– 624
- Ring E, Zobel G. Urinary infection and malformations of urinary tract in infancy. Arch Dis Child.1988;63 :818– 820
- Tsingoglou S, Dickson JA. Lower urinary obstruction in infancy. Arch Dis Child.1972;47 :215– 217
- ↵Greenfeld SP, Wam J. Vesicoureteral reflux: practical aspects of evaluation and management. Pediatr Nephrol.1996;10 :189– 794
- ↵Littlewood JM. Sixty-six infants with urinary tract infection in first month of life. Arch Dis Child.1972;47 :218– 226
- ↵American Academy of Pediatrics, Committee on Quality Improvement and Subcommittee on Urinary Tract Infection. Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children. Pediatrics.1999;103 :843– 852
- ↵Abbott GD. Neonatal bacteriuria: a prospective study of 1460 infants. Br Med J.1972;1 :267– 269
- ↵Drew JH, Acton CM. Radiological findings in newborn infants with urinary infection. Arch Dis Child.1976;51 :628– 630
- ↵Elashoff JD. N Query Advisor Version 1.0. Los Angeles, CA: Statistical Solutions;1995
- ↵Teleform 6.1. San Marcos, CA: Cardiff Software Inc;1999
- ↵SAS Institute. SAS Version 6.12. Cary, NC: SAS Institute;1999
- ↵Bernstein J, Brown AK. Sepsis and jaundice in early infancy. Pediatrics.1962;29 :873– 882
- Ng SH, Rawstrong JR. Urinary tract infections presenting with jaundice. Arch Dis Child.1971;46 :173– 176
- ↵Escobedo MB, Barton LL, Marshall RE, Zarkowsky H. The frequency of jaundice in neonatal bacterial infections. Clin Pediatr.1974;13 :656– 657
- ↵Mora EU, Rodriguez EP. “Hiperbilirrubinemia y sepsis neonatal” [in Spanish]. Pediatria Perinatal.1982;39 :485– 488
- ↵Seeler RA. Hemolysis due to gram-negative urinary tract infection. Birth Defect Orig Artic Ser.1977;13 :425– 431
- ↵Chavalitdhamrong P, Escobedo MB, Barton LL, Zarkowsky H, Marshall RE. Hyperbilirubinemia and bacterial infection in the newborn. Arch Dis Child.1975;50 :652– 654
- ↵Maisels MJ, Kring E. Risk of sepsis in newborns with severe hyperbilirubinemia. Pediatrics.1992;90 :741– 743
- ↵Kivlahan C, James EJP. The natural history of neonatal jaundice. Pediatrics.1984;74 :364– 370
- ↵Thurman WG. Changes in red cell fragility with infection [abstract]. Am J Dis Child.1960;100 :639
- ↵Roelofsen H, Shoemaker B, Bakker C, Ottenhoff R, Jansen P, Oude Elferink, R. Impaired hepatocanalicular organic anion transport in endotoxemic rats. Am J Physiol.1995;269 :427– 434
- ↵Goldman M, Lahat E, Strauss S, et al. Imaging after urinary tract infection in male neonates. Pediatrics.2000;105 :1232– 1235
- Copyright © 2002 by the American Academy of Pediatrics