Posterior urethral valves (PUV) are the most common cause of renal impairment in boys during early childhood. Although antenatal suspicion of this pathology has become quite common in recent years, prenatal diagnosis remains challenging. The aim of this study was to evaluate the predictive value of different ultrasound criteria currently used to diagnose PUV.
We reviewed the antenatal and postnatal files of 54 male patients referred to our center from 2000 to 2006 after detection of fetal bilateral hydronephrosis. The following ultrasound criteria were evaluated in relation to the postnatal diagnosis of PUV: amniotic fluid volume, bladder wall thickness, bladder dilatation and the presence of the ‘keyhole sign’.
Forty-two fetuses (77.8%) were suspected to have PUV on prenatal examination. Out of these, 29 (69.0%) had PUV confirmed postnatally. The sensitivity and specificity of the antenatal diagnosis of PUV were 94% and 43%, respectively. Increased bladder wall thickness and bladder dilatation were highly associated with the diagnosis of PUV (P < 0.001). However, a thick-walled bladder was observed in 39.1% and a dilated bladder in 47.8% of the infants with a postnatal diagnosis other than PUV. The presence of the keyhole sign was not found to predict a diagnosis of PUV (P = 0.27).
Hydronephrosis in a male fetus can be an indicator for several different possible diagnoses, including posterior urethral valves (PUV), vesicoureteral reflux (VUR), pyeloureteral junction stenosis, prune belly syndrome and primary megaureter1–4. The prognosis and appropriate management can differ greatly between these pathologies5–8. It is therefore important to correctly differentiate these various pathologies for parental counseling, to determine in which center delivery should occur and, in centers where prenatal treatment is performed, to adequately select fetuses likely to benefit from a shunting procedure9. The ‘keyhole sign’, representing a dilatation of the posterior urethra in patients with posterior urethral obstruction10, is thought to be a very specific sign of PUV. In the late 1990s, studies assessing the accuracy of prenatal ultrasonography in the detection of PUV3, 4 found a low specificity for PUV diagnosis and contradictory specificities when the different ultrasound signs, including the keyhole sign, were analyzed separately.
The resolution of ultrasound machines has improved in recent years, allowing greater precision in the diagnoses that can be made. The aim of our study was to evaluate the accuracy of various prenatal sonographic criteria for the diagnosis of PUV, with emphasis on the presence of the keyhole sign.
We retrospectively reviewed the antenatal and postnatal records of all patients referred to our center from 2000 to 2006 for bilateral fetal hydronephrosis. The primary outcome of this study was the evaluation of the sensitivity and specificity for the diagnosis of PUV in our center. The capability of different prenatal ultrasound criteria to accurately predict PUV was also assessed. This study was approved by the local ethical committee.
All patients underwent at least one complete ultrasound evaluation performed by one of the five experienced sonologists working in our center. A Voluson 730 ultrasound machine (GE Medical Systems, Zipf, Austria) with a 4–8-MHz transducer was used for all patients. Based on the main prenatal diagnostic suspicion, patients were divided into two groups: (1) patients suspected to have PUV and (2) patients in whom a diagnosis of PUV was not suspected. For each patient it was noted whether or not the keyhole sign had been observed at any point throughout the pregnancy. This sign should correspond to the urethral dilatation between the valves and the bladder neck. The secondary parameters evaluated were: presence of oligohydramnios, gestational age at first diagnosis of oligohydramnios, presence of bladder dilatation and thickness of the bladder. All ultrasound parameters were evaluated subjectively, according to the sonologist's evaluation, and we chose not to review images. We did not perform a vesicoamniotic shunt on any of the patients.
The definitive diagnosis in each case was confirmed by pediatric examination, ultrasound and cystourethrogram when necessary for liveborn children. When parents chose to terminate the pregnancy owing to severe fetal renal impairment, definitive diagnosis was confirmed by anatomopathological examination. Based on the definitive diagnosis the patients were redivided into two further groups: (1) patients who had a confirmed diagnosis of PUV and (2) patients who had pathologies other than PUV diagnosed postnatally.
Results are expressed as frequency and percentage for categorical variables and as mean (± SD) or median (range) for continuous variables. Differences in categorical variables between the two postnatally defined groups were tested using the chi-square test or Fisher's exact test, as appropriate. Differences in continuous variables were tested using Student's t-test or the Wilcoxon rank sum test, as appropriate, and P < 0.05 was considered statistically significant. Statistical analyses were performed using the R software (R Foundation, Vienna, Austria).
Fifty-six male fetuses were referred to our center with bilateral hydronephrosis during the study period. Two patients were excluded from the study because of lack of follow-up or absent anatomopathological diagnosis. Each patient underwent prenatal examination between one and seven times following referral to our center (mean, 5.2). Forty-eight patients (88.9%) were evaluated more than once.
Of the 54 fetuses included in our analysis, 42 (77.8%) were prenatally suspected to have PUV. Out of these, 29 (69.0%) had PUV confirmed postnatally, while 13 (31.0%) had a different diagnosis on postnatal investigation (Table 1). Twelve fetuses were not suspected to have PUV on prenatal evaluation, but two of them (16.7%) were found to have PUV on postnatal investigation. The sensitivity and specificity of our antenatal ultrasound examination for the postnatal diagnosis of PUV were 94% (95% CI, 87–99%) and 43% (95% CI, 30–57%), respectively. Median gestational age at first diagnosis of bilateral hydronephrosis was 25 (range, 19–36) weeks in the fetuses with a final diagnosis of PUV, and 23 (range, 12–33) weeks in those postnatally diagnosed with other pathologies (P = 0.60).
Table 1. Final diagnoses of the 13 patients who were suspected to have posterior urethral valves on prenatal examination but found not to postnatally
Prune belly syndrome
Bilateral junction stenosis
Unilateral ureteral duplication with obstructive ureterocele
Ectopic obstructive ureteral implantation
The keyhole sign was present at least once during fetal life in 51.6% of the fetuses in which PUV was confirmed postnatally. However, it was also present in 34.8% of the fetuses with other postnatal diagnoses (Table 2), and it was not significantly associated with a postnatal diagnosis of PUV (P = 0.27).
Table 2. Presence of keyhole sign in fetuses that were later diagnosed with pathologies other than posterior urethral valves
Patients with keyhole sign (n)
Bilateral junction stenosis
Prune belly syndrome
Ectopic obstructive ureteral implantation
Out of the 48 patients who had more than one ultrasound examination performed at our center, the keyhole sign was persistently present in 11 (22.9%). Three of those (27.3%) did not have a diagnosis of PUV on postnatal evaluation.
The bladder was considered thick walled on prenatal ultrasound examination in 93.5% of fetuses with a postnatal diagnosis of PUV and in 39.1% of fetuses with other pathologies diagnosed postnatally (P < 0.0001). Dilatation of the bladder was present in 96.8% of fetuses with a postnatal diagnosis of PUV and in 47.8% of fetuses without (P < 0.0001).
Oligohydramnios was found to be more frequent in infants with a postnatal diagnosis of PUV (61.3%) than in those who did not have PUV (30.4%) (P = 0.024). In those with a final diagnosis of PUV, the gestational age at first diagnosis of oligohydramnios was significantly lower than in those without PUV (median (range): 33 (20–37) vs. 36 (32–41) weeks, P = 0.008). However, the absence of oligohydramnios did not exclude PUV diagnosis, since 38.7% of fetuses with a normal amniotic fluid index had PUV.
The combination of dilated and thick-walled bladder was observed in 90.3% of fetuses with a final diagnosis of PUV and in 39.1% of fetuses with other pathologies diagnosed postnatally (P = 0.0001). The combined finding of a dilated and thick-walled bladder with the presence of the keyhole sign was not statistically associated with a final diagnosis of PUV. This combination occurred in 45.2% of fetuses with a final diagnosis of PUV and in 26.1% of fetuses with other pathologies diagnosed postnatally (P = 0.18). Table 3 shows the prevalence of each ultrasound parameter in each of the groups defined by final diagnosis and the statistical significance of differences between the groups.
Table 3. Prevalence of prenatal ultrasound signs in those fetuses that were postnatally diagnosed with posterior urethral valves (PUV group, n = 31) and those that were diagnosed with other pathologies (non-PUV group, n = 23)
Dilated and thick-walled bladder with the keyhole sign
In the present study we observed high sensitivity (94%) but low specificity (43%) for the prenatal diagnosis of PUV. Most of the infants followed had urinary tract malformations, such as PUV, major VUR and bilateral pyeloureteral junction stenosis, that required early diagnostic procedures and close neonatal follow-up. High sensitivity for PUV is important to pick up most of the fetuses and neonates with this malformation, since early postnatal evaluation and treatment of the infravesical obstruction are associated with better prognosis. High specificity is equally important. The ability to correctly diagnose fetuses with PUV enables the medical team to provide adequate information to parents regarding prognosis and follow-up after birth. Furthermore, in centers where vesicoamniotic shunts are used, the correct diagnosis of PUV would improve the efficacy of this treatment9. The problem of the low specificity in our center was partly offset by the fact that even when postnatal investigations revealed pathologies other than PUV, most of these neonates benefited from a close follow-up during the neonatal period and infancy at a tertiary center11, 12.
Although few data are available in the literature3, 4, the presence of the keyhole sign is usually considered to be highly predictive of the diagnosis of PUV. Montemarano et al.3 evaluated 21 fetuses presenting with bilateral hydronephrosis. Seventy percent of infants with PUV confirmed postnatally presented dilatation of the posterior urethra on prenatal ultrasound examination. They also found that 50% of patients with prune belly syndrome presented this sign transiently during fetal life. In that study, dilatation of the posterior urethra was found to be very specific for the diagnosis of PUV. In the same year, Abbot et al.4 evaluated 22 fetuses presenting with bilateral hydronephrosis and persistent megacystis. In contrast to the study performed by Montemarano et al., only 25% of fetuses in which the keyhole sign was observed had PUV confirmed postnatally. Seventy-five percent of fetuses presenting with this sign had other pathologies diagnosed postnatally, such as VUR.
In our series, the keyhole sign was not a reliable predictor of the diagnosis of PUV, as this sign was present at least once during prenatal evaluation in 35% of fetuses presenting pathologies other than PUV after birth, including 50% of the fetuses who had reflux diagnosed postnatally. The reason for this low specificity must be determined. Some degree of hypotonicity of the vesical neck and dysynergic vesical contraction during urodynamics has been described in 30% of boys presenting with VUR13. This hypotonicity is usually seen on voiding cystourethrograms as a dilatation of the vesical neck, an image that could simulate a dilatation of the posterior urethra on prenatal ultrasonography (Figure 1). Indeed, in voiding cystourethrograms, the presence of PUV is usually seen as a dilatation of the posterior urethra above the valves, downstream of the vesical neck (Figure 2). This finding was detected on prenatal ultrasound for certain fetuses that were diagnosed with valves postnatally, and this finding could be more specific for PUV diagnosis. However, this observation may be transient and only detectable during bladder emptying (Figure 3). Cohen et al.14 published a case report in which the same finding was observed in a patient later diagnosed with PUV. Further studies are required to evaluate the accuracy of this finding for the diagnosis of PUV.
Unlike the keyhole sign, bladder dilatation and wall thickness are thought to be highly predictive of PUV diagnosis. In accordance with this, we found that the presence of a dilated or thick-walled bladder correlated highly with the diagnosis of PUV. However, at least one of these signs was also present in 48% of patients with other postnatal diagnoses. Other studies evaluating the role of these variables in PUV diagnosis have shown divergent results. Abbot et al.4 and Maizels et al.15 showed that the presence of a thick-walled and dilated bladder during fetal life could be inidicative of several diagnoses, including PUV, VUR, and megacystis/megaureter. On the other hand, Montemarano et al.3 found no bladder dilatation or increased wall thickness in fetuses that had pathologies other than PUV diagnosed postnatally. In neonates, dilated and thick-walled bladders can be associated with diseases other than PUV, especially VUR. Yeung et al.16 evaluated 117 male children presenting with VUR who had prenatal hydronephrosis, and showed that these children had some degree of thickness (56%) and enlargement (34%) of the bladder. Likewise, Avni and Schulman17 found a thickened bladder in 32% of newborns diagnosed with VUR. Thus, both the present and previous studies show that fetuses with VUR may show an enlarged or thickened bladder. Furthermore, as the bladder is a distensible organ, bladder wall thickness is highly dependent on its emptying. Some studies have evaluated an index correlating bladder wall thickness and bladder length for the prediction of urinary obstruction in children18, 19. These studies have revealed significantly higher indices in cases with bladder obstruction. The same may be present in fetuses, and bladder evaluation of thickness/length index associated with additional dynamic evaluation of the urethra during voiding may provide clues for the more accurate prediction of the diagnosis of PUV.
This was a retrospective study. We chose not to re-evaluate ultrasound images since they were all performed and interpreted by experienced sonologists in an everyday practice setting. Sensitivity for the prenatal diagnosis of PUV is high, but despite improvements in ultrasound technology, our study has shown that specificity for the diagnosis of PUV remains low.
We would like to thank M. C. Aubry, J. P. Aubry, B. Barthe, S. Chemouny and C. Morere for their contribution to this study. We would also like to thank CAPES and COFECUB for supporting Brazilian–French international collaboration.