Assessing Bladder Volumes in Young Children Prior to Instrumentation: Accuracy of an Automated Ultrasound Device Compared to Real-time Ultrasound

Abstract

ACADEMIC EMERGENCY MEDICINE 2011; 18:816–821 © 2011 by the Society for Academic Emergency Medicine

Abstract

Objectives:  Automated bladder ultrasound (ABUS) devices are portable and designed to provide automated measurement of bladder volume. They are simple and require minimal training compared to conventional real-time ultrasound (RTUS). Their most common application in the acute pediatric setting is to assess bladder volumes prior to performing invasive urine collection such as suprapubic aspiration (SPA) in children younger than 2 years of age. However, data on ABUS in young children are limited. The aim of this study was to assess the repeatability and accuracy of one type of ABUS, the BladderScan, in measuring of bladder volume in children aged 0 to 24 months when compared with RTUS.

Methods:  Healthy children aged 24 months and younger were scanned twice, 1 hour apart, using ABUS and RTUS. ABUS readings were performed by two senior pediatric emergency physicians who both completed three readings for each child. The measurements were repeated using a second ABUS machine in case of machine variability. RTUS measurements were performed by a pediatric sonographer who was blinded to the ABUS results. ABUS and RTUS measurements were compared by Bland-Altman analysis to determine the repeatability coefficient (repeatability) and the limits of clinical agreement (accuracy).

Results:  Bladder volume measurements were performed on 61 children aged 0 to 24 months (31 males; mean ± SD = age 11 ± 6.2 months; range = 0 to 24 months) using both the ABUS and the RTUS. There was wide variation between ABUS and RTUS measurements. The repeatability coefficient within ABUS readings was 20 mL. By Bland-Altman analysis, the 95% limits of agreement between ABUS and RTUS were –31 to +19 mL. ABUS also detected no values between 0 and10 mL.

Conclusions:  This study showed poor repeatability and accuracy in bladder volume measurements using BladderScan ABUS when compared to RTUS. The ABUS method does not appear to be a reliable method for assessing bladder volumes in children aged 0 to 24 months prior to bladder instrumentation.

Urinary tract infection (UTI) is present in 4.1% to 7.5% of children aged 0 to 24 months who have fever with no identifiable source.^1–3 Delay in the diagnosis and treatment of UTI in young children increases the risk of septicemia and permanent renal scarring.^4–7 Recommendations on urine collection methods to rule out UTI in febrile young children vary, with the U.K. National Institute of Clinical Excellence recommending clean catch,⁸ and the American Academy of Pediatrics recommending bladder instrumentation with suprapubic aspiration (SPA) or urinary catheterization as best methods of urine collection to rule out UTI.⁹ SPA is considered the criterion standard for sterile urine collection and has the lowest rate of contamination by fecal and skin flora.^10–15

There are few risks associated with SPA;^16,17 however, SPA is painful,^18,19 and blind attempts (i.e., without imaging) have reported success rates of approximately 50%.^20–27 This is often due to insufficient urine in the bladder at the time of the procedure. Repeated attempts at SPA expose children to unnecessary pain and may increase the slight risks of hematuria, gross hemorrhage, bowel perforation, peritonitis, abdominal wall abscess, and bacteremia.^17,28–35

The use of conventional real-time ultrasound (RTUS) to confirm bladder fullness prior to SPA can increase the rate of successful SPA to 79% to 100%;^{20,22,24,26,27,36,37} however, RTUS machines are expensive and require advanced training. Automated bladder ultrasound devices (ABUSs) are portable ultrasound devices designed for automated calculation of bladder volume. ABUS offers an attractive alternative to RTUS, as the devices are less expensive and require minimal training. Previous studies have shown similar usability for experienced and learner users of ABUS³⁸ and no significant learning curve over an 8-month study period.³⁹ However, data on the accuracy of ABUS in children aged 0 to 24 months are limited in terms of comparisons with bladder volume determined by catheterized volumes or RTUS.^40,41 Other studies^40–44 assessing the accuracy of ABUS in children up to 18 years may not apply to children below 2 years of age, for whom the bladder is smaller and still largely an intraabdominal organ. To the best of our knowledge, there are no studies assessing the accuracy of a recent model ABUS, the BladderScan BVI 9400 (Verathon, Bothel, WA) in children.

At our large tertiary pediatric institution, the BladderScan ABUS is routinely used in the emergency department and on wards prior to bladder instrumentation in young children. Anecdotally, the accuracy of the readings seemed problematic; in particular, repeated readings on the same patients appeared to produce variable results. We therefore set out to 1) determine the repeatability of this device for bladder volume measurements in children aged less than 2 years and 2) examine the accuracy of the ABUS, with RTUS by a pediatric sonographer as the criterion standard.

Methods

Study Design

This was a prospective observational study. The Royal Children’s Hospital (RCH) Human Ethics Research Committee approved the protocol (HREC #29052A). All parents were consented prior to their participation in the study. Once they were recruited an information sheet was given to them and on the day of the study they signed written consent prior to their participation. The written consents were all collected and kept in secure files. No parents withdrew their consent.

Study Setting and Population

This study was conducted in the radiology department of the RCH in Melbourne, Australia, between August 2009 and October 2009. Healthy children aged less than 24 months were recruited through advertisements posted to nursing and medical staff of RCH. We excluded children with a history of renal tract abnormalities, abdominal surgery, or abdominal scar tissue; open skin wounds; or wounds in the suprapubic area.

Study Protocol

The study procedure was divided into two time points: initial bladder volume measurements using both machines (t = 0) and second bladder volume measurements using both machines (t = 1; Figure 1). At t = 0, the bladder volume was measured using both conventional RTUS (ACUSON S2000, Siemens, Erlangen, Germany) and the Verathon BladderScan BVI 9400, the most current model ABUS available. Two pediatric emergency consultants each performed sets of three readings on each child using the ABUS. Measurements were performed as outlined in the manufacturer’s manual: the probe was placed over the lower abdomen, with adequate gel applied. The probe was repositioned until all arrows were flashing, indicating that the probe was positioned over the bladder. Three readings were obtained by pressing the scan button and reading the volume in milliliters off the screen. An additional ABUS of the same type (“BladderScan B”) was added to the study for the final 36 children to ensure that results were not due to any individual ABUS variability. This was motivated by the apparent variability with the initial ABUS device (“BladderScan A”). All following ABUS measurements were performed using both machines. Both machines had been calibrated to manufacturer’s standards before inclusion in the study. ABUS measurements were immediately followed by RTUS measurement. All RTUS measurements were performed by a single senior pediatric sonographer employed at RCH who was blinded to the previous ABUS readings. Bladder volume was calculated using a standard ellipsoid formula that has been previously validated in the literature: length × width × height × 0.52375.⁴⁵ All RTUS views and measurements were saved to the machine for verification.

Figure 1.  Study flow comparing real time ultrasound and automated ultrasound measurements.

The child left the radiology room for an interval period of 1 hour and returned for a second bladder volume measurement at t = 1. Children were permitted to eat or drink as normal during the 1-hour interval period. The purpose of this was to increase the dataset by capturing the bladder of each child at two independent volumes. Measurements were repeated using both the ABUS and the RTUS as described above for the initial bladder volume measurement.

Data Analysis

The data were analyzed by the technique described by Bland and Altman for measuring limits of agreement and the repeatability coefficient in method comparison studies.^46,47 The Bland Altman limits of agreement are the margins within which 95% of the differences between the two methods lie. The narrower the limits (i.e., the closer they are to 0) the better the agreement between the two methods. The Bland-Altman repeatability coefficient is defined as twice the standard deviation of the differences between repeated readings. This means that an ideal repeatability coefficient would be 0, indicating that there was no difference in repeated measurements. For example, if the repeatability coefficient is 20 mL, this can be interpreted to mean that 95% of the time, repeated measurements on the same subject at the same point in time were within 20 mL of each other.

Data were descriptively analyzed using EpiData v3.1 (EpiData Association, Odense, Denmark) and STATA 10 (StataCorp, College Station, TX). We also stratified data and analysis post hoc by age (0 to less than 6 months, 6 months to less than 12 months, and 12 months to less than 24 months).

Results

Bladder volume measurements were performed on a total of 61 children (31 males; mean age ± SD 11 ± 6.2 months, range = 0–24 months), each at two separate time points. All 61 children had a RTUS measurement at each time point (122 RTUS measurements). All 61 children had six ABUS measurements at each time point with bladder scanner A.

The final 36 children (of the 61) had an additional six ABUS measurements at each time point with bladder scanner B. Corresponding ABUS (BladderScan A only) and RTUS measurements for each child are presented in Figure 2. The data in Figure 2 are sorted by the bladder volume measured by RTUS (crosses) with corresponding ABUS measurements (circles) shown in a vertical line above and below RTUS readings. Poor repeatability within repeated ABUS measurements at each time point was seen, as indicated by the range of ABUS measurements on each vertical line. Poor agreement between ABUS and RTUS measurements was shown by the difference between RTUS measurements and their corresponding ABUS measurements. In addition, the ABUS did not record any readings between 0 and 10 mL.

Figure 2.  Corresponding ABUS (BladderScan A only) and RTUS measurements for each child. ABUS = automated bladder ultrasound; RTUS = real-time ultrasound.

The repeatability coefficients were similar for both BladderScan A (19 mL) and BladderScan B (17 mL). Considering the similarity between machines, further analysis was performed using the measurements from both BladderScan A and BladderScan B together. The overall repeatability coefficient was 20 mL, indicating that 95% of repeated ABUS measurements were within 20 mL of each other (Table 1).

Table 1.    Repeatability and Accuracy of the BladderScan Ultrasound Device by Age

	N = 61	Repeatability (Repeatability Coefficient)	Accuracy (95% Limits of Agreement)	Range of Difference Between ABUS and RTUS
ABUS = automated bladder ultrasound; RTUS = real-time ultrasound.
0 to <6 months	18	20 mL	–25 to +14 mL	–40 to +32 mL
6 to <12 months	19	17 mL	–38 to +17 mL	–60 to +42 mL
12 to < 24 months	24	23 mL	–28 to +23 mL	–56 to +36 mL
Total	61	20 mL	–31 to 19 mL	–60 to 42 mL

The accuracy (95% limits of agreement) after adjusting for repeated measures was –31 to +19 mL. This can be interpreted to mean that 95% of ABUS readings were between –31 and +19 mL of the corresponding RTUS measurement. The range of difference between ABUS and RTUS measurements was –60 to 42 mL. The measurements as divided into the three age groups are shown in Table 1. The repeatability and accuracy of ABUS readings were similar across all age groups.

Children with known renal tract abnormalities were excluded from the study; however, the sonographer identified a possible case of bladder wall thickening in one child and possible vaginal distension in a second child. Both children were referred for follow-up ultrasound investigation. Follow-up investigation confirmed bladder wall thickening in the first child, which was diagnosed as posterior urethral valves requiring surgery, but ruled out vaginal distension in the second child.

Discussion

We assessed the utility of an ABUS device in assessing bladder volumes in young children. We found that the repeatability of the BladderScan device, i.e., the ability to obtain consistent repeated readings in the same patient at the same time point, is poor. Further, we found that when compared with bladder volumes determined by RTUS, the accuracy of ABUS was also poor. Although our data only apply to the model studied, the utility in the acute care setting of any model ABUS irrespective of design without documented repeatability and accuracy data in children under 2 years of age is problematic for clinical use.

Data on accuracy and repeatability of ABUS devices in young children are limited. To our knowledge, there are no published validation data for the ABUS model used in our study. In an earlier BladderScan model (BVI 3000), Rosseland et al.⁴¹ found a tendency to underestimate bladder volume in children below 3 years of age when compared to catheterized volumes, but still recommended the ABUS as a valuable aid in avoiding postoperative overdistension of the bladder based on a mean difference of –18 mL (95% confidence limits –60 to 20 mL). De Gennaro et al.⁴⁰ assessed the accuracy of the BladderScan BVI 2500 when compared to RTUS in 41 children aged below 3 years and found the machine to be clinically unacceptable in young children based on a mean difference of –6.5 mL (95% confidence limits –35.3 to 12.3 mL).

We noted that the ABUS did not identify any volumes less than 10 mL. We can only assume that the ABUS is not sensitive enough to pick up such low volumes.

While the use of conventional RTUS to assess bladder volume prior to SPA has been shown to increase success rates,^{20,22,24,26,27,36,37} there are no clinical data to show an increased success rate using ABUS devices, except from one study that used a different type of automated ultrasound device (Bardscan, Chatswood, Australia). Munir et al.²³ performed three repeated BardScan measurements on children prior to attempting SPA and found that SPA was successful 79% (31/39) of the time when a minimum reading of 10 mL with no zero readings was obtained, and 100% (24/24) of the time when a minimum volume of 20 mL without zero readings was obtained.²³

Although seemingly attractive from the perspective of lower cost and less operator training required compared to RTUS, we found that the ABUS model we investigated had poor repeatability and accuracy when assessing bladder volumes in young children. ABUS use should be assessed for accuracy in young children with small bladders before deployment. Unless ABUS applicability can be shown in young children, RTUS appears to be the best tool to determine bladder size prior to instrumentation for sterile urine specimens in the acute pediatric setting.

Limitations

Real-time ultrasound was used as the criterion standard. Optimally, the bladder volume would have been determined using bladder catheterization, which was not ethically possible in well children. In addition, both RTUS and ABUS devices are operator-dependent. However, the RTUS operator was an experienced pediatric sonographer at a major pediatric radiology department, and all bladder measurements were digitally recorded for verification. While RTUS bladder volume was calculated using a standard validated formula,⁴⁵ a number of different methods have been proposed to calculate bladder volume.⁴⁸ ABUS readings were performed according to the instructions of the manufacturer and recorded independently by two emergency consultants. We also used two identical machines to guard against a malfunctioning device. However, both devices provided identical accuracy and reliability outcomes. A further study we are currently conducting will assess both RTUS and ABUS volume measurements prior to performing SPA, to compare the success rates between the two different methods.

Conclusions

Our study showed poor repeatability and accuracy in bladder volume measurements using BladderScan automated bladder ultrasound when compared to real-time ultrasound. The automated bladder ultrasound studied does not appear to be a reliable method for assessing bladder volumes in children aged 0 to 24 months prior to bladder instrumentation.

Acknowledgments

We acknowledge the children and families who supported this study. We acknowledge grant support from the Murdoch Children’s Research Institute, Melbourne, Australia.