Introduction

The Randomized Intervention for Children with Vesicoureteral Reflux (RIVUR) trial was a multisite, randomized, placebo-controlled trial aimed at evaluating the efficacy of antimicrobial prophylaxis in preventing recurrences of urinary tract infection (UTI) in children with vesicoureteral reflux (VUR) diagnosed after an index UTI [1]. A total of 607 children 2 to 71 months of age were randomized in the ratio of 1:1 to receive either antimicrobial prophylaxis or placebo daily and were followed for 2 years. The study found that long-term antimicrobial prophylaxis reduced the risk of UTI recurrences by 50% (23.6% vs. 12.9%, corresponding to an observed number needed to treat (NNT) of 10) [2]. Because the use of long-term antimicrobial prophylaxis likely leads to the development of antibiotic resistance and alterations of the microbiome [3] and because the NNT observed in the RIVUR trial was relatively large, routine use of long-term antimicrobial prophylaxis for all grades of VUR remains controversial.

Accordingly, there is an interest in identifying higher-risk subgroups of children that would benefit the most from long-term antimicrobial prophylaxis. Two previous studies have used patient-level data from the RIVUR trial to examine subgroups of children for whom long-term use of antimicrobial prophylaxis could be advocated. Wang et al. (2018) used the data from the RIVUR study to test the performance of a model they had previously developed using data from a relatively small study conducted in a referral population [4]. Shaikh et al. (2020) used a cost-utility model to evaluate the trade-off between benefits and risks of prophylaxis based on VUR grade and concluded that treating children with grade IV VUR is cost-effective [5]. While helpful in clarifying the clinical and financial trade-offs in children with VUR, the latter study did not explicitly take into account covariates other than the grade of VUR.

In this manuscript, we sought to further explore data from the RIVUR trial to identify subgroups of children with VUR who appear to benefit the most from long-term antimicrobial prophylaxis. We used penalized logistic regression methods to systematically identify subgroups that consider both the treatment effect in the subgroup and the size of the subgroup.

Methods

Analytic sample

The study design and data collection of the RIVUR trial has been described previously [1]. In the RIVUR trial, 39 of 302 children (event rate, 0.13) had a recurrent UTI in the antibiotic prophylaxis treatment arm compared to 72 of 305 children (event rate, 0.24) in the placebo control arm. Five children with missing VUR grade were excluded from the analysis, leaving an analytic sample of 602 children. A total of 38 children in the treatment arm and 42 in the control arm were lost to follow-up before their final visit (2 years after enrollment) and did not have a UTI recurrence while in the trial. Because 486 of 607 children were followed for at least 1 year without a recurrence, and because most observed recurrences occurred within the first year (the interval between enrollment and a 10% incidence of recurrence was 336 days and 106 days in the treatment arm and the control arm, respectively) [2], we analyze these participants as without an event following the example of the original analysis undertaken in the trial.

Covariates

Baseline demographic and clinical characteristics of the RIVUR trial participants have been reported previously [1]. For this subgroup analysis, we include three covariates assessed at enrollment (baseline) that had a p-value for interaction with treatment of less than 0.12: index UTI type (febrile or symptomatic), categorical grade of VUR (I–IV), and bladder and bowel dysfunction (BBD). BBD was determined by administering the dysfunctional voiding symptom scale [6] to parents of toilet-trained children. Because BBD was not assessed in children 2–23 months of age, we create a composite covariate that combined age and BBD. This composite covariate has four levels: (1) age 2–23 months and not toilet trained, (2) age 24–71 months and not toilet trained, (3) age 24–71 months and toilet trained with BBD absent, and (4) age 24–71 months and toilet trained with BBD present. Risk factors used in creating our model have been identified in previous studies [7,8,9,10].

Statistical methods

We generated a pool of candidate subgroups through thresholding linear predictors of fitted penalized logistic regression models with treatment, covariates, covariate-by-covariate interactions, and treatment-by-covariate interactions [11]. This approach, employing penalized regression, does not stratify the sample but rather analyzes it as a complete dataset. This methodology provides a robust framework for effectively handling complex, relatively small datasets, significantly reducing the risk of overfitting which is often a challenge in more traditional regression models. Furthermore, this approach may lead to candidates that are mixtures of distinct subgroups because it evaluates the effects of various factors collectively rather than in isolation. We supplemented the pool by considering clinically relevant single-covariate subgroups. For each candidate subgroup, we calculated the observed NNT. More information on model fitting, variable selection methods, and generation of candidate subgroups is available in the Supplementary Materials.

Of note, we required all candidate subgroups containing children with low-grade VUR to also contain children with higher-grade VUR, because, from a clinical perspective, if children with low-grade VUR benefit from prophylaxis, so would those with higher-grade VUR. Similarly, we required all candidate subgroups including toilet-trained children without BBD to also include toilet-trained children with BBD. We performed all statistical analyses using R [12].

Results

Our analytic sample includes a total of 602 children with a median age of 12 months. These children are categorized into four baseline BBD–age composite categories, as follows: 400 children aged 2–23 months who were not toilet trained, 76 children aged 24–71 months who were not toilet trained, 55 children aged 24–71 months who were toilet trained and without BBD, and 71 children aged 24–71 months who were toilet trained and with BBD present. Regarding the categorical grade of VUR at baseline, our sample distribution is as follows: 68 children had grade I, 254 had grade II, 230 had grade III, and 50 had grade IV. In terms of the index UTI type at baseline, 516 children had febrile UTI, whereas 86 experienced a symptomatic UTI. It is noteworthy that our analytic sample had 5 fewer children than the main study, which included 607 children. The 5 excluded children had missing VUR grade.

Characteristics of top subgroups (i.e., those with an observed risk difference greater than or equal to 0.2), identified and ordered by observed risk difference, are presented in Table 1. The subgroup of children with grade IV VUR, BBD, and febrile index UTI (observed risk difference = 0.5, NNT = 2) had the largest observed risk difference among all candidates but was very small, representing less than 1% of the children in the RIVUR sample. The second subgroup included children with BBD and febrile index UTI (7% of the RIVUR sample; observed risk difference = 0.48, NNT = 3). The third subgroup included children with BBD (12% of the RIVUR sample; observed risk difference = 0.33, NNT = 4). The last subgroup included children with grade IV VUR (8% of the RIVUR sample; observed risk difference = 0.2, NNT = 5). It is noteworthy that all top subgroups fall within the broader category of children with BBD (and any grade of VUR) and those with grade IV VUR (regardless of BBD status), which also emerged as a statistically identified candidate subgroup; however, its composition and intermediate risk difference do not provide additional clinical insights beyond what is already provided by the top subgroups. Overall, the total sample size for these top subgroups is 117, comprising 19% of the RIVUR sample with an observed risk difference of 0.27 and a NNT of 4. Figure 1 plots the identified top subgroups and other candidate subgroups.

Table 1 Characteristics of the top subgroups (i.e., those with an observed risk difference greater than or equal to 0.2). These four subgroups appeared to benefit more from long-term antimicrobial prophylaxis than other subgroups
Full size table
Fig. 1
figure 1

Candidate subgroups with a positive observed risk difference are plotted with dot symbols. The top four candidate subgroups (A to D), each with an observed risk difference greater than or equal to 0.2, are labeled (see Table 1 for subgroup details). The candidate subgroup of children with BBD (and any grade of VUR) and those with grade IV VUR (regardless of BBD status) is labeled with “C + D,” indicating that it is a combination of C and D. A horizontal line at 0.11 represents the observed risk difference among all study participants. Other clinically relevant single-covariate subgroups are also labeled

Full size image

In summary, antimicrobial prophylaxis appears to be particularly relevant for children with BBD (and any grade of VUR) and those with grade IV VUR (regardless of BBD status). This comprised 117/602 (19% of the RIVUR sample) of children included (Fig. 2). Children with a febrile UTI who did not have grade IV VUR or BBD did not appear to benefit from antimicrobial prophylaxis (71% of the RIVUR sample, observed risk difference of 0.09, NNT of 12) and neither did children with grade III VUR but no BBD (34% of the RIVUR sample, observed risk difference of 0.03, NNT of 35).

Fig. 2
figure 2

Venn diagram illustrating subgroups of children (with counts) in the RIVUR trial. Children in the shaded subgroups, which together comprise 19% of the RIVUR sample, appear to benefit more from long-term antimicrobial prophylaxis than others

Full size image

Discussion

We found that BBD appears to have been driving the differences in the risk of recurrent UTI in the RIVUR trial. In addition to children with BBD, children with grade IV VUR (regardless of BBD status) also appear to benefit from long-term antimicrobial prophylaxis.

Our results are generally consistent with previous studies but provide a clearer picture of the subgroups that may benefit from treatment. A previous cost-effectiveness analysis considering only grade of VUR as the only covariate found that long-term antimicrobial prophylaxis appeared to be most cost-effective in children with grade IV VUR [5]. While Wang et al. [4] suggested that females and uncircumcised males with both grade I–III VUR and BBD are also at high risk, these risk factors were selected based on data from a relatively small study performed in a referral population, not patient-level data from the RIVUR study [13]. Kent and Hayward have noted that the treatment effect is usually higher in subgroups with higher risk [14]. This was true in our data; the top subgroups we identified also had high rates of breakthrough UTIs.

What do our results mean for the practitioner? Shaikh et al. [5] suggested that the treatment of children with grades IV was clearly cost-effective. While our results also support the treatment of children with grade IV VUR, we found that, in addition to children with grade IV VUR, those with lower grades of VUR who have BBD also appear to derive substantial benefit from antimicrobial prophylaxis. Both this study and the study by Shaikh et al. suggest that the treatment of children with grade III VUR who do not have BBD might not be particularly effective.

There are several limitations in our investigation. First, the RIVUR trial was not powered to detect treatment effect heterogeneity [2]. Second, there were very few children with grade IV VUR (8.3% of the RIVUR sample). Third, the RIVUR study excluded children with grade V VUR. Fourth, our results are only directly applicable to children similar to those enrolled in the RIVUR trial [15]. Because changes in imaging following the 2011 American Academy of Pediatrics guideline on the management of young children with UTI [16] may have changed the distribution of children with various grades of VUR, our results may not be directly applicable to children currently diagnosed with VUR. Fifth, rather than overall utility, which would include both efficacy and harm, we focused on assessing differences in efficacy. Sixth, for most children enrolled, details regarding treatment received for baseline BBD were not recorded. Thus, we are not able to assess the effect of treatment for BBD on our results. We did so because the trial lacked details to support in-depth analyses of harms related to antibiotic use [2]. Lastly, the assessment of BBD remains controversial. Our results are based on the methods used in the RIVUR trial in which the dysfunctional voiding symptoms scale [6] was used to identify children ≥ 24 months of age with BBD; more work on standardizing definitions of BBD across studies is needed.

In conclusion, our re-analysis of the RIVUR trial data suggests that the use of long-term antimicrobial prophylaxis appears to be most beneficial for children with BBD (and any grade of VUR) and children with grade IV VUR (regardless of BBD status). While we have identified subgroups that appear to benefit more than others from the use of antimicrobial prophylaxis, on average, 4 children still need to be treated with 2 years of antimicrobial prophylaxis to prevent one additional UTI. Given new data on the harms associated with long-term antimicrobial prophylaxis use [3], it is therefore not clear whether routine treatment of children in the identified subgroups is justified. Further study is needed to clarify the risks and benefits of treatment in these subgroups.