Who Needs an Ultrasound? Using Patient Symptom Questionnaire & UTI History to Better Inform the Decision to Obtain RBUS in Children with Non-neurogenic Lower Urinary Tract Dysfunction
Katherine M. Fischer, MD1, Amanda Berry, PhD, MSN, BSN, CPNP1, Adriana Messina, MSN, CRNP, FNP-BC1, Ethan Samet, MD2, Stephen Zderic, MD1, Jason Van Batavia, MD1.
1Children's Hospital of Philadelphia, Philadelphia, PA, USA, 2Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA.
Background Non-neurogenic lower urinary tract dysfunction (LUTD) affects up to 20% of children. However, there is no consensus as to the initial LUTD work-up including the use of imaging like RBUS. Predicating which patients would benefit from RBUS is challenging as the majority will demonstrate normal imaging. We hypothesized that higher DVISS scores and a positive urinary tract infection (UTI) history would be associated with a higher likelihood of RBUS abnormality and may be useful in guiding this decision.
Methods We retrospectively reviewed new clinic patients from 5/2014-1/2016 to identify those seen for LUTD who received an RBUS as part of their work-up. RBUS results were divided into two groups 1) clinically significant abnormalities and 2) normal and insignificant findings (ex-simple renal cyst). Demographic data, DVISS score at initial visit and UTI history were abstracted on chart review. Patients with neurogenic bladder or known urologic abnormalities were excluded. Association between positive UTI history, DVISS score, gender, and race and RBUS abnormality were evaluated using univariate and multivariate logistic regression analysis. Receiver operating curves (ROC) curves were created to evaluate the predictive model and a Youden index calculated to determine the optimal cutoff for DVISS score to predict abnormal RBUS.
Results We identified 333 patients who received an RBUS as part of LUTD work-up and 15 (4.5%) had a clinically significant abnormality. Median age was 9 years and 56.8% were female (Table 1). Significantly more patients with abnormal RBUS had a positive UTI history (p = 0.019) and median DVISS was significantly higher (p=0.002). Both positive UTI history and DVISS score were associated with RBUS abnormality on univariate analysis (OR=3.61, 95%CI=1.26-10.34; p=0.017 and OR=1.11, 95%CI=1.04-1.20; p=0.003) whereas neither gender nor race were. Positive UTI history and DVISS score were significant on multivariate analysis as well (OR=4.74, 95%CI=1.55-14.5; p=0.006 and OR=1.13, 95%CI=1.05-1.22; p=0.001). A ROC curve for this model had AUC=0.8 (Figure 1A). A DVISS score cutoff of 12 was determined to be ideal for predicting abnormal RBUS with AUC=0.68, sensitivity=80% and specificity=57% (Figure 1B). Using a cutoff of DVISS>12 and positive UTI history, patients with both risk factors were significantly more likely to have an abnormal RBUS than those with zero or one risk factor (p=0.041) (Table 2).
Conclusions We attempted to identify factors associated with increased likelihood of significant RBUS abnormality in pediatric LUTD patients and create an evidence-based approach to determining who should undergo imaging. We found that DVISS score>12 and UTI history are useful in determining when to obtain an RBUS in this population.
| Normal RBUS (n=318) | Significant Abnormal RBUS (n=15) | p-value | |
| Age (yrs), median (IQR) | 9 (7, 12) | 8 (5, 14) | 0.41 |
| Race | 0.65 | ||
| Asian | 5 (1.6%) | 0 (0.0%) | |
| Black | 50 (15.7%) | 4 (26.7%) | |
| Indian | 1 (0.3%) | 0 (0.0%) | |
| Other | 61 (19.2%) | 3 (20.0%) | |
| Pacific Islander | 1 (0.3%) | 0 (0.0%) | |
| Unknown | 2 (0.6%) | 0 (0.0%) | |
| White | 198 (62.3%) | 8 (53.3%) | |
| Female | 178 (56.0%) | 11 (73.3%) | 0.29 |
| Positive UTI History | 62 (19.5%) | 7 (46.7%) | 0.019 |
| DOVE Score, median (IQR) | 11 (6, 17) | 17 (13, 25) | 0.002 |
| Risk Factors | Normal or Insignificant Finding | Abnormal & Significant Finding |
| None | 138 (100%) | 0 |
| UTI only | 35 (92.1%) | 3 (7.9%) |
| DVSS >12 only | 118 (93.7%) | 8 (6.3%) |
| UTI & DVSS >12 | 27 (87.1%) | 4 (12.9%) |
Back to 2022 Abstracts