National Trends in Voiding Cystourethrograms During Hospitalization for Young Infants With Urinary Tract Infections

We performed a retrospective cohort study of infants aged 0 to 2 months hospitalized with a UTI at a Pediatric Health Information System (PHIS) participating institution from January 1, 2008, to December 31, 2019. The PHIS database contains administrative data from 47 children’s hospitals across the United States. After excluding 9 hospitals that did not have complete data for the entire 12-year study period, data from 38 hospitals were included. We evaluated trends in the performance of VCUG over time, as well as outcomes including recurrent UTI, diagnosis of VUR, and antiurinary reflux procedures performed. This study was approved by the institutional review board at the study institution with a waiver of informed consent.

Infants from aged birth to 2 months of age discharged from either an inpatient or observation unit with a primary diagnosis of UTI were included. For infants with more than 1 hospitalization for UTI, only the initial hospitalization was included and was considered the index hospitalization. UTI was defined by International Classification of Disease, Ninth Revision (ICD-9) of 5990, 77182, and 59010 and by International Classification of Disease, Tenth Revision (ICD-10) of N390, P393, and N10. We excluded subjects with preexisting urologic conditions (Supplemental Table 3).

We assessed rates of VCUG performance, as well as rates of recurrent UTI, diagnosis of vesicoureteral reflux, and antiurinary reflux procedures performed at the index institution. VCUG performed during the index hospitalization was identified using imaging service codes. (Supplemental Table 3). A recurrent UTI diagnosis was considered when an infant had either an emergency department (ED) visit or hospitalization with a primary diagnosis of UTI from 10 days to 1 year after the index hospitalization. We did not include visits occurring before 10 days because these visits would likely fall within the treatment period for the initial UTI based on median treatment duration for neonatal UTI.¹⁸  We defined a VUR diagnosis as an ED visit or hospitalization with a diagnosis code (in any position) associated with VUR (Supplemental Table 3) within 1 year of index hospitalization. Last, we defined the performance of an antiurinary reflux procedure as any encounter with a current procedural terminology, ICD-9, or ICD-10 code associated with a procedure to correct VUR (Supplemental Table 3) within 2 years of index hospitalization.

The study period was divided into 2 discrete periods: preguideline (2008–2011) and postguideline (2012–2019). We described patient characteristics using medians and interquartile ranges for continuous variables and frequencies and percentages for categorical variables. The trend over time of the binary outcomes of VCUG performance during index hospitalization, recurrent UTI within 1 year, VUR diagnosis within 1 year, and performance of an antiurinary reflux procedure within 2 years were compared between preguideline and postguideline periods using a piecewise mixed-effects logistic regression model. To account for clustering by hospital, random intercepts for hospitals were included. We further adjusted for age, sex, insurance, urban residence, season, complex condition, ICU hospitalization, length of stay, lumbar puncture, complete blood count, and chemistry. Odds ratios of each outcome for 1-year increase in time were estimated with 95% confidence intervals (CIs) and P values. Pre-post difference was evaluated by setting a breakpoint at the end of year 2011 and testing the difference in the odds ratios for 1 year before and after the breakpoint.

A total of 17 578 infants aged 0 to 2 months were hospitalized with primary diagnosis of UTI over the 12-year study period. A total of 391 infants met 1 or more exclusion criteria and 17 187 infants were included in this study (Fig 1).

The characteristics of the study cohort are shown in Table 1. Over time, there were statistically significant differences in patient age, urban residence, primary insurer, and season of visit. Over the course of the study period, median length of stay decreased, fewer children had an ICU stay, and the proportion of children with a chronic condition increased. Fewer patients had lumbar punctures, complete blood counts, and chemistry testing performed when comparing the postguideline and preguideline groups.

The percentage of patients who had VCUG performed during index hospitalization decreased over the study period (Fig 2). The number of infants and proportion of infants for whom VCUG was performed at each hospital are shown in Fig 3. VCUG performance ranged from 2% to 73% across hospitals, and there was no correlation between the number of infants hospitalized with UTI at each hospital and VCUG performance (Pearson correlation coefficient 0.28, P = .09). Trends in the following occurring at the index hospital are presented in Fig 4: recurrent UTI within 1 year, VUR diagnosis within 1 year, and performance of an antiurinary reflux procedure within 2 years.

In the preguideline period, the odds of VCUG performance decreased by 21% per year (adjusted odds ratio [OR], 0.79; 95% CI, 0.77–0.81; P < .001); in the postguideline period, the odds of VCUG performance decreased by 20% per year (adjusted OR, 0.80; 95% CI, 0.77–0.83; P < .001). This preguideline and postguideline difference is not statistically significant (P = .60) (Table 2). In the preguideline period, the odds of recurrent UTI diagnosis within 1 year increased by 11% per year (OR, 1.11; 95% CI, 1.08–1.15; P < .0001). However, over the postguideline period, there was no statistically significant change in the odds of UTI diagnosis within 1 year (P = .07). This preguideline and postguideline difference is statistically significant (P < .001). Although this preguideline and postguideline difference is statistically significant, this difference relates to a stabilization of the slope after 2014 (Supplemental Table 4). The odds of VUR diagnosis within 1 year decreased by 5% per year in the preguideline period (adjusted OR, 0.95; 95% CI, 0.92–0.97; P < .001) and by 8% per year in the postguideline period (adjusted OR, 0.92; 95% CI, 0.88–0.95; P < .001). The OR for VUR diagnosis within 1 year was not statistically significant when comparing the preguideline and postguideline periods. The odds of an antiurinary reflux procedure being performed decreased by 15% per year in the postguideline period (adjusted OR, 0.85; 95% CI, 0.78–0.92; P < .001), and this was a statistically significant difference from the preguideline period (P = .02).

In this study, we observed that rates of VCUG performance during index hospitalization steadily declined among infants aged less than 2 months hospitalized with UTI. However, there was not a significant change in the rate of decline after the publication of the 2011 UTI Clinical Practice Guideline that recommended limiting the performance of VCUG for infants aged 2 to 24 months. The lack of change in slope may be reflective of the fact that the initial decrease in VCUG performance began before publication of the guideline.

We observed decreased VCUG performance during index hospitalization in young infants is similar in the preguideline and postguideline periods. This suggests that the observed shift in practice away from performing VCUG during index hospitalization for these young infants is likely multifactorial. Previous work has demonstrated that duration of intravenous antibiotics for young infants with UTI has been decreasing,²¹  and, as a result, we observed a decrease in length of stay for these young infants. This decrease in length of stay may reduce the opportunity to obtain VCUGs during hospitalization. Additionally, there is a national trend toward performing diagnostic studies in the outpatient setting.²⁰ 

The question remains whether the decline in VCUG performance during index hospitalization in the youngest infants represents a shift to performing VCUG after discharge or decreased VCUG performance overall. Over the past 2 decades, evidence has accumulated that questions the traditional teaching that performing a VCUG soon after diagnosis of UTI posed a risk of medical harm to the patient. Studies have demonstrated that VCUG early in the treatment course does not lead to worsening UTI or increase risk of VUR diagnosis because of ureteral dilation during acute infection.^18,19,22  Additionally, multiple studies suggest that VCUG is less likely to be performed when recommended after inpatient discharge.^17,18 

We found that the rate of recurrent UTI within 1 year increased in the preguideline period, but the rate remained steady in the postguideline period, despite a continued decline in the rate of VCUG performance. This argues that the slightly increased rate of recurrent UTI is not directly related the decline in VCUG performance during index hospitalization for these infants and may be reflective of other factors such as increased regionalization of pediatric care.²³ 

Because VCUG performance during index hospitalization decreased over the study period, we found a decrease in VUR diagnoses within a year and a decrease in antiurinary reflux procedures in the 2 years following the index hospitalization. Although it is likely that our follow-up period does not fully capture these outcomes, our findings are consistent with previous research demonstrating decreased incidence of VUR in infants aged 2 to 24 months and a decrease in antiurinary reflux procedures in all children since 2011.^3,4  Our findings of decreased VUR diagnoses suggest that fewer VCUGs are being performed in total rather than simply a shift to the outpatient setting.

Future work should determine to what extent VCUG performance in these young infants has decreased versus shifted to the outpatient setting. To fully assess the risks and benefits of delaying VCUG in infants aged 0 to 2 months, future efforts should focus on determining the risk of long-term complications including renal scarring, chronic kidney disease, and hypertension. Also, delay in VCUG could result in delayed diagnosis of urologic abnormalities, such as posterior urethral valves. Therefore, studies investigating different approaches to VCUG performance in this population specifically are warranted.

Our data have several limitations. Because the data source used an administrative database for ED and inpatient encounters, we do not have access to information from outpatient visits or from institutions not participating in PHIS, nor do we have access to clinical information, culture, or VCUG results to confirm the diagnoses. We included young infants with first hospitalization for UTI at the index hospital, but we are unable to determine if they had a prior UTI elsewhere, had fever with this UTI, or confirm culture results. We included infants transferred from outside institutions, which likely results in an underestimation of the prevalence of laboratory testing, including urinalysis. We are limited to investigating VCUGs performed during inpatient encounters and we are unable to determine if VCUGs have been performed in the outpatient setting. Our outcome data for recurrent UTI, VUR diagnoses, and antiurinary reflux procedures captures children who had subsequent inpatient or ED visit to the same institution and for which the billing code for UTI was used. Additionally, we are unable to account for reflux severity or grading given limitations of the administrative database used in this study. We are unable to capture outpatient visits for UTI or confirm that the patient truly had a recurrent UTI. Similarly, we chose to define antiurinary reflux procedures using codes for procedures that could be indicated for correction of VUR, and we cannot definitively determine if the true indication for the procedure was VUR. We are unable to see results of testing, such as renal ultrasound results and thus we cannot verify diagnoses or comment on how providers decide to obtain VCUG. We chose exclusion criteria based on previous published research³  and therefore we did not exclude all possible preexisting urologic abnormalities. In addition, our data are limited to children’s hospitals that care for a higher proportion of children with comorbidities and higher acuity illnesses. Last, our observation of trends over time in UTI and VUR diagnosis, and of antiurinary reflux procedures do not imply a causal relationship because other temporal factors cannot be accounted for.

Rates of VCUG among infants aged less than 2 months hospitalized with UTI have declined steadily over the 12-year study period without a change in the rate of decline preguideline and postguideline period. The rate of recurrent UTI diagnoses increased in the preguideline period but remained steady in the postguideline period. Rates of vesicoureteral reflux diagnoses and performance of antiurinary reflux procedures have decreased with a more significant decline in the postguideline period. Future research is needed to determine the total rate of VCUG performance including those performed in the outpatient setting. Additionally, efforts should explore whether reduction in VCUG performance during hospitalization among young infants hospitalized with UTI has an effect on rates of renal scarring, hypertension, and chronic kidney disease.