Accuracy of postvoid residual volumes after vaginal delivery: a prospective equivalence study to compare an automatic scanning device with transurethral catheterization

After delivery, abnormal postvoid residual volumes (PVRV) after spontaneous voiding attempts are encountered in up to 45% of women [1‐5]. This condition is also known as asymptomatic or covert postpartum urinary retention (PUR). Although evidence for long-term clinical consequences of abnormal PVRV postpartum is not available, literature suggests that even a single episode of bladder overdistension can result in prolonged micturition problems, urinary tract infections (UTIs), and renal obstruction [6‐9]. Therefore, accurate and timely diagnosis is important.

To measure PVRV, different techniques are available. Transurethral catheterization is considered the gold standard for accurate measurement of bladder volume and PVRV. However, real-time ultrasonography (US) has proven to be a valid and reliable substitute [10, 11]. Nevertheless, the use of a regular US device requires specific expertise and training and a mathematic calculation after measurement. This makes the method unsuitable for everyday clinical use. In order to facilitate the recognition of urinary retention, automatic portable devices have been developed. In many fields of medicine, screening for PVRV with a portable BladderScan® device is part of routine care, for example, in patients with neurogenic bladder, voiding dysfunction, or after orthopedic or urogynecological surgery [12‐14]. In this way, unnecessary transurethral catheterizations can possibly be reduced [15, 16]. The effect of such a measure could be the reduction of nosocomial infections, patient discomfort, and healthcare costs. However, the technique is not widely accepted for postpartum women. A reason for this could be that the portable bladder scanner might not only measure bladder volume but also the (remaining bloody content of the) uterus and its pronounced vascularization. While the reliability of the BladderScan® is confirmed in men and nonpregnant women [17‐19], its reliability in the puerperium remains inconclusive [20, 21]. Although several studies have compared portable bladder scanning devices with catheterization, these studies were either small or validation was done for relatively high residuals [4, 22‐25]. To determine the accuracy of the BladderScan® in women directly after vaginal delivery, a prospective equivalence study was conducted.

Between June 2012 and May 2017, this study was conducted in three Dutch teaching hospitals after approval by the medical committee of the Catharina Hospital in Eindhoven, The Netherlands (M11–1165), followed by the two other centers.

Women were asked to participate in this study prior to their delivery, if they were ≥ 18 years old and were to deliver vaginally. Women were either counselled and included before the onset of labor or were asked to participate after having given birth. Written informed consent was obtained from all patients.

Data of 407 patients could be used for final analysis. Table 1 shows patient characteristics.

Median volume of catheterization measurements was 375 ml (0–1800 ml), the 25th percentile was 150 ml, and the 75th percentile was 600 ml. The median volume of the BladderScan ® measurements was 380 ml (0–999 ml), with the 25th and 75th percentile being 200 ml and 567 ml, respectively. Mean difference between measurements was −13 ml (SD ± 177 ml). These results are presented in Table 2.

Agreement between measurement methods are presented in Fig. 1.

Spearman’s rank correlation coefficient calculation showed a close correlation between BladderScan® and catheterized bladder volumes (r = 0.82, p < 0.01). Spearman’s rank correlation coefficient was repeated after excluding extreme data (i.e., BladderScan > 999 ml only and BladderScan > 999 ml plus catheterization volume >1000 ml.) With these restrictions, outcomes remained similar, resulting in close correlations of r = 0.80 and r = 0.78, respectively. An agreement plot with 95% confidence interval (CI) showing mean average of each measurement compared with the difference between measurements is shown in Fig. 2. Here, a good agreement over the whole measurement range of 0–1000 ml was found.

Table 3 shows mean PVRV by BladderScan® and catheterization and mean difference between the two categorized per 100-ml group (as measured with a BladderScan®).

Table 4 shows sensitivity and specificity calculated using a PVRV ≥500 ml as a cutoff value. PPV and NPV were 91.9% and 75.3%, respectively.

Based on our study results, we conclude that the BladderScan® is an accurate and therefore valid alternative to the invasive measurements obtained by catheterization in measuring PVRV directly after vaginal delivery. By measuring PVRVs in a large and nonselected group of women after vaginal delivery, a high degree of agreement was found between the two measurement methods. This agreement was found in the whole range of PVRVs, which is known to be very wide directly postpartum [5, 27]. We believe this is sufficient evidence to support the use of the BladderScan® BVI-9400 after vaginal delivery. Measuring PVRV is not part of standard postpartum care, and various authors have shown that increased residual volumes often normalize within several days [27, 28]. However, in daily clinical practice, there is a need to identify abnormal PVRV—for instance, when micturition produces very small portions of urine.

To minimize the risk of damage to the bladder, accurate and timely measurement of PVRV is important to determine whether treatment should be started. In the absence of other techniques, and because of a slight advantage that catheterization is a measurement technique as well as a treatment, it has been the accepted gold standard for PVRV measurements. While it can cause considerable physical discomfort and inevitably increases the risk for UTIs, noninvasive alternatives are nevertheless preferable.

Evidence regarding accuracy of noninvasive methods after delivery is scarce. However, concerning automatic scanning devices, seven earlier studies reported their validity or reliability [4, 20‐25]. Most of these studies reported on a small patient sample and in selected cases [21, 24, 29]. Our study had a considerably larger sample size and was prospectively designed as an equivalence study to objectively assess accuracy of the BladderScan®. Further, this study was conducted in an unselected cohort of women, whereas other studies were potentially biased by including women at risk of having abnormal PVRV, who were only catheterized in case PVRV >300 ml, or with an indwelling Foley catheter in situ, which was clamped to measure bladder volumes [21, 24, 25]. This inevitably leads to selection bias and therefore validation of the BladderScan® only in women with higher residual volumes.

Furthermore, PVRV of patients in our study were measured using the same, newer, model, i.e. BVI 9400. To our knowledge, this study is the first to determine the accuracy of this newer model. Previous models (BVI 3000 and BVI 6100) were, according to the manufacturer, inferior in discriminating bladder volumes from the uterus itself.

A recurring thought is that an automatic scanning device could overestimate residual volumes, as it would also measure the content of the uterus. However, others demonstrated by repetitive US after delivery that in the early puerperium, the cavity of the uterus is generally empty, and fluid and debris are mostly seen in the middle puerperium (from the 7th day after delivery) [30]. In addition, median volume of the uterus after delivery is 22 cm³. This shows that the thoughts on overestimation of PVRV by the BladderScan® after delivery are not based on evidence and are therefore unfounded [31].

Nevertheless, there are some limitations to our study that need to be addressed. First, a large group of nurses performed the measurements. Despite the ease of use and the structured training they all received before the study commenced, this still may have introduced variation in measurements. While we have no data on measurements recorded by different persons measuring for this study, interobserver coefficients cannot be reproduced. Potentially, with a smaller nursing staff or more extensive training, performance of the BladderScan ® would have been better. However, despite this, we believe measurements done by a larger group of nurses reflects daily practice. Further, the effect of an interobserver variation is expected to be small, as a very strong correlation was found between measurements. The second limitation is that the time between measurement with the BladderScan® and catheterization was often not recorded in patient files. Both transurethral catheterization and scanning will only produce reliable values when measurements are performed directly after micturition to minimize refilling due to continuing diuresis [32, 33]. Although by protocol measurements had to be performed directly after each other, we do not have data indicating whether the protocol was followed in all cases. This would have been especially interesting in cases in which extreme differences were found between the two techniques, possibly resulting in a better correlation.

The importance of any measurement technique is that it is accurate in discriminating between patients who need to be treated and who do not. We have shown that this type of scanning device is suitable for all volume ranges. However, to minimize overtreatment, a valid cutoff value to discriminate physiology from pathology is essential. In the literature, asymptomatic (covert) PUR is defined as a PVRV ≥150 ml after spontaneous micturition. Previously, we suggested increasing this cutoff value to ≥500 ml based on the high residual volumes found in a previous unselected cohort of women after vaginal delivery [5]. With a median PVRV of 375–380 ml found in this study (±216–315 ml), we believe that the commonly used limit of 150 ml lies well within the physiological range and probably results in unnecessary management of a physiological phenomenon. Using the new cutoff value of PVRV ≥500 ml, we believe the BladderScan® can accurately be used in clinical daily practice, with a specificity of 85.4% and sensitivity of 85.6%. (Table 4).

The BladderScan® BVI 9400 is an accurate alternative to evaluate PVRV after vaginal delivery and is preferred over transurethral catheterization to reduce the risk of UTI and optimize patient comfort.