INTRODUCTION

Urinary tract infection (UTI) is a common condition in infancy, and is commonly associated to urinary tract abnormalities, particularly to the vesicoureteral reflux (VUR) existing in 30%-40% of infants with UTI.1,2

Indication of imaging studies in infants with UTIs is controversial. The most recent guidelines of the American Association of Pediatrics2 recommend performing a SVCU in all infants under 2 years of age with UTI, but widely different opinions exist about what subsequent examinations are indicated and when they should be done.3-9

In this setting, a formula has recently been proposed to predict the presence of VUR and avoid unnecessary SVCUs.10 Other authors have questioned the validity of this formula.11 The purpose of our study was to assess this formula and propose alternatives to it.

PATIENTS AND METHODS

Infants diagnosed of their first urinary tract infection (UTI) at the infant section, either in the outpatient clinic or the hospital ward, from January 2005 to June 2007 were studied.

All infants having positive urine cultures, CRP values at diagnosis, an ultrasonography performed during the UTI episode, and a serial voiding cystoureterography (SVCU) within 1 month of UTI diagnosis were enrolled into the study. Fifteen infants were excluded from the study due to lack of any of the variables analyzed.

Risk was calculated using the Oostenbrink score (score = 6 * male gender + 7 * positive family history ¿ 1 * age + 1*CRP + 14 for dilation in renal ultrasonography).10

Sensitivity, specificity, and positive and negative odds ratios, with their 95% CIs, were also calculated using the statistical page 2-way Contingency Table Analysis available in the Internet.12

A multivariate study was performed by binary logistic regression using commercial SPSS 11.0© software,13 starting from the maximum model including all variables in the Oostenbrink formula plus the presence of urine cultures positive for pathogens other than E. coli. Variables not achieving a significance value of p < 0.05 were removed from the model one by one.

RESULTS

Atotal of 267 infants were enrolled into the study. Age ranged from 2 days to 24 months, with a mean of 4.2 months (standard deviation 4.79 months, median 2.50 months). One hundred and sixty infants (59.5%) were males. The clinical history showed that 4 infants (1.5%) had a family history of VUR, and 137 (51.3%) were being breastfed, or had been breastfed for at least 3 months.

Pathogens responsible for UTIs were divided into two groups, E. coli (82.67%) and non E. coli (17.33%). The most common pathogen in the latter group was Klebsiella (22 cases), followed by enterococci (14 cases), Proteus, Citrobacter, and Streptococcus agalactiae.

A total of 40 infants (14.98%) had abnormalities in the kidneys and/or urinary tract in ultrasonography, in most cases dilations of the pyelocalycial and pyeloureteral system (pyelic and ureteral dilations were defined as values greater than 18 mm and 10 mm in the transverse axis and distal ureter respectively), and 108 infants had a pathological SVCU (40.44%), with a total of 175 kidneys with VUR. The probability of VUR was calculated using the Oostenbrink formula.10 Table 1 shows the cases with scores greater and lower than 0 (the advocated cut-off point) and VUR. As shown, this value was exceeded in most cases and only 14 infants had this value, 8 of whom had VUR.

Table 2 gives the distribution of VUR presence and grade according to the International Study Classification14 (the maximum grade was considered in bilateral VURs), depending on whether an abnormal ultrasonography and/or UTI by pathogens other than E. coli had been seen, and Table 3 shows the results of logistic regression for the presence of VUR and grade < 3 VUR. Statistical significance was only found with the presence of abnormalities in ultrasonography and non-E. coli UTI in both cases.

Table 3 shows the sensitivity, specificity, and positive and negative odds ratios for the presence of non-E. coli urine cultures and/or ultrasonographic abnormalities for the presence of VUR, grade < 3 VUR, and grade < 4 VUR. The negative odds ratio was 0 for VUR < 4 and 0.25 for VUR < 3.

DISCUSSION

The cumulative incidence of UTIs in children under 6 years of age has been estimated at 3%-7% in females and 1%-2% in males. Imaging tests have been recommended after the first UTI to assess VUR presence and grade. SVCU has an essential place among these tests because it allows for identifying the presence of VUR and its grade.2-4,6 Therapeutic decisions deriving from the presence of VUR are related to its grade and include surgical correction, prophylactic treatment, and follow-up without treatment. A wide controversy exists about the most appropriate treatment, but most recent studies appear to suggest that the course is similar in patients with VUR regardless of whether or not they receive prophylactic treatment, at least in cases with less severe VUR (grades I-II), which are a majority.15-17

VUR is found in 30%-40% of children with UTIs.2 The proportion of children with VUR in our study was 40.44%. In order to avoid performance of SVCU, a traumatic examination exposing to a high amount of radiation, in the  60%-70% of children showing no VUR, Oostenbrink proposed use of a formula that considered age, family history, dilations in renal ultrasonography, and CRP values. Using this formula, only one case of VUR was excluded from diagnosis in his study.10 Other authors have questioned this formula, pointing out that it only had a 3% specificity for any type of VUR, and a 13% specificity for grade < 3 VUR.11 Application of the Oostenbrink formula in our study was shown to be of little help, first because it would only exclude performance of SVCU in 12 cases (5.24%), but also because VUR was found in 8 of these patients.

The presence of UTIs caused by pathogens other than E. coli has also been reported to be a risk factor for the presence of urinary tract abnormalities, including VUR (OR, 8.16; 95% CI, 1.88-35.48).18 This was also seen in our study for the presence of VUR (OR, 3.98; 95% CI, 1.99-7.95).

In the calculation performed using logistic regression for the presence of both any VUR and grade < 3 VUR, both the variables in the Oostenbrink formula and the existence of an urine culture revealing a pathogen other than E. coli were introduced. The result showed that only the presence of abnormalities in ultrasonography and the existence of positive non-E. coli urine cultures maintained statistical significance. All grade < 4 VURs, 80.6% of grade < 3 VURs, and 57.47% of all VURs were detected under these conditions, but it should be noted that only four cases with grades higher than 4 were found.

It should also be noted that the need for detecting grade I-II VURs has been questioned because of their good course without treatment. The proposed indication for SVCU in all cases of UTI is therefore not clear.2,15-17,19

The main limitation of our study was that it only enrolled children under 24 months of age referred to a hospital, who are supposed to have the most severe symptoms, though most infants with a febrile syndrome are usually referred for hospital diagnosis and management.

The conclusion drawn from our study is that it appears reasonable to advocate that SVCU is only performed in the first UTI episode in children under 2 years of age having abnormalities in renal ultrasonography and/or urine cultures revealing pathogens other than E. coli. In any case, the conduct of larger studies to confirm these findings would be appropriate.