How may Doppler indices help in the differentiation of obstructive from nonobstructive hydronephrosis?

BACKGROUND: We assess the potency of different Doppler indices in the differentiation of obstructive and nonobstructive hydronephrosis.
MATERIALS AND METHODS: In this study, infants and children who were referred for the evaluation of unilateral hydronephrosis were enrolled. Ultrasonography for the assessment of the degree of hydronephrosis and a voiding cystourethrogram for the exclusion of vesicoureteral reflux was performed. Then, Doppler ultrasonography was done for both kidneys of each patient using four classic Doppler indices as well as the difference (delta) of each index between to kidneys. Diuretic renography with 99 mTc-ethylene dicysteine (99 mTc-EC) was performed for each patient.
RESULTS: Thirty-nine patients met the inclusion criteria. After diuretic renography, 29 (74.35%) patients had shown a nonobstructive pattern, and ten (25.65%) patients had a partial (intermediate) or complete obstruction. Using receiver operating characteristic (ROC) curve, none of the classic indices of Doppler duplex (i.e., resistive index [RI], resistance index, end diastolic velocity, and peak systolic velocity) had the ability to make a difference between obstructive and nonobstructive hydronephrosis. However, by calculating the difference (delta) of these indices between two kidneys of each patient, delta RI could differentiate the nonobstructive condition, significantly (P = 0.006). A cutoff value of 0.055 has 60% sensitivity and 82.8% specificity. The area under the ROC curve for delta RI is 0.795 (standard error: 0.086, 95% confidence interval [CI]: 0.626, 0.964). Furthermore, RI ratio between two kidneys of each patient could differentiate the nonobstructive condition, significantly (P = 0.012). A cutoff point of 1.075 has 70% sensitivity and 82.8% specificity. The area under the ROC curve for RI ratio was 0.769 (standard error: 0.104, 95% CI: 0.565, 0.973).
CONCLUSION: This study shows that RI ratio and delta RI with a high specificity could differentiate nonobstructive hydronephrosis and therefore it is a promising way to use especially in the follow-up of children with hydronephrosis.

INTRODUCTION

Nowadays, imaging modalities have a crucial role in the diagnosis and follow-up of the urinary tract disorders in pediatrics. The widespread use of prenatal ultrasonography offers the early detection of some of these disorders and provides a guide for further evaluation.[1] Ureteropelvic junction obstruction (UPJO) is the most common cause of upper urinary tract obstruction in children.[2] In this disorder, lack of normal urine passage from the renal pelvis to ureter is considered, mainly because of a functional insufficiency. The increased pressure of renal pelvis due to the accumulation of obstructed urine may result in a progressive decrease in renal function. Ultrasonography is a beneficial tool in the prenatal as well as postnatal detection of hydronephrotic kidneys.[3] Furthermore, it has been used in the monitoring of the alteration in the degree of hydronephrosis, assessment of cortical thinning or scar, and exclusion of other obstructive etiologies in children with hydronephrosis. However, ultrasonography fails to provide the essential information about the physiologic and functional state of the kidney which is required to confirm the diagnosis of UPJO.[4567]

Nuclear renogram has become the standard modality for the assessment of renal function and urinary flow in children with a primary diagnosis of UPJO. This study is usually delayed until the 2nd month after birth to allow for stabilization of glomerular filtration rate and often is performed after a voiding cystourethrogram (VCUG) excludes vesicoureteral reflux (VUR).[89] Diagnosis of UPJO by nuclear renography may be based on decreased differential renal function (<40%) or a reduction in washout time of radioisotope agent.[10] Besides this controversy on the definition of UPJO which causes intraobserver and interobserver variability in different centers, nuclear renography has its own false negative and positive results. In fact, UPJO is neither uniform nor straightforward diagnosis, even by applying multiple modalities include nuclear renography. Indications of follow-up nuclear renography in known cases of UPJO are another subject of debate in pediatric uroradiology.[10]

Duplex Doppler ultrasound is a tool which can provide useful information about hemodynamics of kidneys while it is nonionizing modality. It has been demonstrated that obstruction causes a decrease in vascular flow due to increasing in vascular resistance.[111213] Resistive index (RI) is the most valuable duplex index used in the obstructive renal states. Diagnostic value of RI has been studied in the evaluation of hydronephrosis in children as well as in adults.[1415] In children with unilateral obstruction, increase of RI in intraparenchymal arteries may be a remarkable finding.[16] However, RI is a nonspecific index and increased renal RI may be seen in urinary obstruction, parenchymal injury, and also vascular abnormality. Moreover, RI is an age-dependent parameter, and applying the adult's threshold value in young children can be misleading.[1718]

It has been shown that the utility of RI ratio and a comparison of RI values between the kidneys of each patient are more reliable than using an RI cutoff value in the evaluation of children with unilateral obstruction.[18]

In this study, we assess the potency of different duplex indices to find their value in the differentiation of obstructive and nonobstructive hydronephrosis. This study may provide an adjunct strategy to resolve controversial renogram results or to decrease the frequency of renograms in the follow-up of known cases of UPJO.

MATERIALS AND METHODS

Patient selection

In a prospective study between January 2014 and March 2015, children aged older than 2 months who were referred by pediatric urologist or nephrologist to the Radiology Department for evaluation of unilateral hydronephrosis were enrolled, consecutively. Primary diagnosis of UPJO was considered using the gray-scale US in the patients with unilateral hydronephrosis (anteroposterior diameter of pelvis >7 mm), beaking of UPJ, and normal diameter of ureter without evidence of pelvicalyceal calculus. All patients underwent a standard VCUG[19] for the exclusion of VUR or any probable anatomical malformation. In this study, we applied the normal contralateral kidney as a control for each case to compare indices. Therefore, patients with bilateral UPJO, horseshoe kidney, or any other developmental abnormalities (including multicystic dysplastic kidney, contralateral VUR, megaureter, retrocaval ureter, or vascular malformations) were excluded from the study. Furthermore, the cases with long-standing severe UPJO in which no parenchyma for detecting vessels was available were excluded from the study. The ethics committee approved this study, and written informed consent was obtained from all subjects.

Doppler duplex study

All infants and children were well fed or hydrated before the survey. The Doppler US was done during slow breathing and sometimes under sedation with oral administration of 0.5–1 cc/kg chloral hydrate. Both gray-scale and Doppler examination were performed using a Voluson E8 ultrasound machine (GE Healthcare), by the use of the linear and curved transducers. For localization of vessels, color Doppler with a 2–4 mm Doppler gate was used. In the initial evaluation by color Doppler, any evidence of bridging artery, microaneurysm, stenosis, or thrombosis in the course of both renal arteries from their origin to arcuate branches was excluded.

All waveforms were measured on the lowest velocity scale (to maximize the waveform sizes) and achieved without obscuring background noise. The angle between the ultrasound beam and the direction of blood flow was adjusted as close as possible to 0°. The average of highest and lowest velocity of three similar consecutive waveforms was selected as peak systolic velocity (PSV) and end diastolic velocity (EDV).

Pulsatility index defined as PSV minus EDV divided by mean velocity. RI was defined as PSV minus EDV divided by PSV. It was measured manually and also by the automatic option of the ultrasound setting in the arcuate and interlobular branches of the upper and lower poles and interpolar parenchyma. Waveforms were obtained at least three times for each kidney and values from these waveforms averaged to obtain mean values for each kidney. These steps were all performed for both kidneys of each patient. Then, the difference between Doppler indices of kidneys of each patient was calculated and recorded. RI ratio was defined as the ratio of RI in the hydronephrotic kidney to that in the healthy contralateral kidney.

Diuretic renography

If the patient had a diuretic renography in recent 3 months, he/she was exempted from a repeat of the renography. In other children, diuretic renography was offered and only the renography which was performed within 1 month since Doppler ultrasonography was accepted and other patients were excluded from the study. In those with new diuretic renography before the Doppler study, the radiologist remained unaware of renography results during ultrasonography to avoid probable bias.

All the diuretic renography was performed in a university hospital center and interpreted by a specialist in the nuclear medicine who was unaware of the Doppler results.

Infants aged under 1 year were well fed by milk 15–30 min before the scan. Other children were orally hydrated with up to 500 ml of water. The parents were asked to ensure about voiding of the child just before the start of the test. The diuretic renography was performed using a Philips (ADAC) Forte Dual Head Gamma Camera equipped with low-energy all-purpose collimators. After the intravenous injection of 4.2 MBq/kg (0.12 mCi/kg) of 99 mTc-ethylene dicysteine (99 mTc-EC) (with a maximum dose of 555 MBq) image acquisition was started immediately. Dynamic acquisition at 64 × 64 matrix was acquired 2 s per frame for 1 min and 1 min per frame for 30 min, from posterior view in the supine position. At 15th frame, the furosemide was intravenously injected (1 mg/kg, the maximum dose of 20 mg). We did not insert a Foley catheter into the bladder, unless occasionally in the noncooperative children. Postvoiding static 1 min images were acquired to quantify excretion.

After grouping the dynamic images, a region of interest over each kidney and the urinary collecting systems was drawn manually. The background radiation was subtracted. Using renogram time-activity curves on the images, obstruction was defined. Clearance half-life <10 min was defined as nonobstructive, 10–19 min defined as intermediate (partial), and >20 min was defined as the obstructive value finding.

Statistical analysis

All analyses were performed in IBM SPSS Statistics (version 20, IBM, Somers, NY, USA). We used receiver operating characteristic (ROC) curve for obtaining the cutoff points. Demographic findings were reported by descriptive tests (mean and standard deviation). For comparing the mean values between two groups, the independent student t-teat was used. P < 0.05 was considered statistically significant.

RESULTS

Of 51 patients who met the inclusion criteria, 39 patients were willing to complete the offered workup and their diuretic renography was performed within the expected time. Twenty-six (66.7%) were male and 13 (33.3%) were female. Their age was ranged from 2 months to 13 years. Twenty (51.3%) were under 2 years and 34 (87.2%) were under 8 years. Figure 1 shows the age distribution of the patients. Hydronephrosis in 32 (82.1%) cases was in left kidney, and 7 (17.9%) was in right kidney. After diuretic renography, 29 (74.35%) patients had shown a nonobstructive pattern (12.82%), 5 patients had an intermediate (partial) obstruction (12.82%), and 5 patients showed a complete obstructive pattern.

Figure 1

Potency of resistive index ratio for diagnosis of obstructive hydronephrosis

Table 1 demonstrates Doppler indices in both nonobstructive and obstructive cases. Using ROC curve, none of the classic indices of Doppler duplex (i.e., resistive index [RI], pulsatility index [PI], EDV, and PSV) had the ability to make a difference between obstructive and nonobstructive hydronephrosis. However, by calculating delta of these indices between two kidneys of each patient, delta RI could differentiate the nonobstructive condition, significantly (P = 0.006). A cutoff value of 0.055 has 60% sensitivity and 82.8% specificity. The area under the ROC curve for delta RI is 0.795 (standard error: 0.086, 95% confidence interval [CI]: 0.626, 0.964).

Table 1

Renal Doppler indices in patients with hydronephrosis

Furthermore, RI ratio between two kidneys of each patient could differentiate the nonobstructive condition, significantly (P = 0.012). A cutoff point of 1.075 has 70% sensitivity and 82.8% specificity. The area under the ROC curve for RI ratio was 0.769 (standard error: 0.104, 95% CI: 0.565, 0.973) [Figures 2 and 3] [Table 2].

Figure 2

Potency of delta resistive index for diagnosis of obstructive hydronephrosis

Figure 3

Age distribution of the patients

Table 2

Valuable renal Doppler indices based on receiver operating characteristic curves for diagnosis of obstructive hydronephrosis

DISCUSSION

In this study, we assess the potency of Doppler indices in the differentiation of obstructive from nonobstructive hydronephrosis in children. Our results show that RI ratio and delta RI with a high specificity could differentiate nonobstructive hydronephrosis, and therefore, it is a promising way to use in the follow-up of children with hydronephrosis.

In our practice, a large number of children with the typical gray-scale appearance of UPJO have normal urine passage in diuretic renography. Moreover, in the follow-up of known cases of stable noncomplete UPJO, interobserver variation about the grading of hydronephrosis sometimes leads to a repeat of a diuretic renography, in the absence of true change in the patient's condition. Another issue is permanent dilatation of pyelocaliceal system in some UPJO children after pyeloplasty, despite the removal of the obstruction. These repetitive conflicts rise the rate of inopportune diuretic renography, as well as health expenditure and parental concern.

Doppler ultrasound has been introduced as an available, nonionizing, low-cost modality in the assessment of some urinary tract diseases in pediatrics. During the past decades, several studies have been done to find the association of alterations in renal arterial waveform and indices with renal parenchymal injury, urinary obstruction, VUR, and vascular disorders.[111320] Renal RI has been introduced as an angle-independent index for quantifying the changes in blood flow that may occur with different parenchymal or collecting system pathologies.[14] Based on the most of the studies, a standard mean renal RI is about 0.60. However, 0.70 is the acceptable upper limit of normal RI in adults. In children at the early 4 years of life, it is frequently seen that mean RI may go more than 0.70.[2122]

In this study, we measured other Doppler indices further than RI, such as PI which is valuable in Doppler study of some splanchnic arteries. However, none of them could differentiate obstructive from nonobstructive conditions.

The efficacy of Doppler sonography in the management of partial UPJO is restricted. Chen et al.[23] illustrated that RI is often elevated at complete or high-grade UPJO, but most cases with incomplete or low-grade UPJO had normal RIs. Furthermore, Cole et al.[2425] in two separate animal experimental studies in pig and rabbit models demonstrated the limitation of Doppler sonography to detect low-grade UPJO.

Several studies have shown that the sensitivity of Doppler sonography in partial UPJO may improve by doing the study after force diuresis (diuretic Doppler sonography).[2627] Although this modality is not approved by general radiology or urology committees yet, recent studies illustrate a technical support of diuretic Doppler sonography in the evaluation of UPJO. The high false-negative rate of this technique may be due to low-grade or partial UPJO, extremely early obstruction or forniceal rupture. In these cases, with severe chronic obstruction, arterial stretchability would be affected because interstitial pressures are normal.

In a study by Riccabona et al.,[28] children with complete obstruction of UPJ showed a significantly increased and asymmetric mean RI which returned to normal value after the operation. In their study, other noncomplete UPJO cases who have not an indication for surgery on diuretic renography showed a normal and symmetrical mean RI. In our study, RI could not differentiate obstructive from nonobstructive cases.

Lim et al. in a study evaluated the usefulness of the RI ratio in differentiation between obstructive or nonobstructive upper urinary tract dilation in children. They demonstrated that an RI ratio (which compares the RI of the dilated kidney with the RI of the normal, contralateral kidney) cut-off value of ≥1.10 appears to be a useful parameter for evaluation and follow-up of unilateral obstructive hydronephrosis in children.[18] In our study, RI ratio of ≤1.075 could differentiate nonobstructive hydronephrosis with 82.8% specificity.

In a study by Lee et al., RI and delta RI had limitations in the diagnosis of unilateral obstructive uropathy due to very low sensitivity. They declared that elevated RI might indicate an acute and severe obstructive uropathy; nevertheless, it is a poor indicator of chronic or partial obstruction.[29]

In the present study, we used 99 mTc-EC as radioisotope agent. 99 mcTc-EC has been used as a substitute for Tc DTPA with apparently more accurate results than that obtained from 99 mTc-DTPA.[30]

Limitations regarding this study should be implied. Some of our patients did not desire to continue the workup or discontinue the follow-up, so our initial sample size decreased. The study patients show a relatively wide age range. In our series, there was a small percentage of older children with recently detected hydronephrosis which were referred for workup and enrolled in the study. Although it was preferable only to enroll the infants, the concentration of patients’ age is in early childhood, as shown in Figure 1. Moreover, considering the use of both the indices and their delta between two kidneys, the effect of age on our results seems not to be significant.

In this study, we used contralateral kidney as a control in each patient, and we measured RI ratio and delta RI for overcoming the potential bias of RI include of dependency to age group and systemic conditions. Our results show that both RI ratio and delta RI are reliable indices with equal value in the diagnosis of nonobstructive hydronephrosis. However, in children with any urinary tract disease or malformation in contralateral kidney, these two indices may have questionable value.

CONCLUSION

Diuretic renography is used to diagnosis of renal function and urinary obstruction, but the interpretation of its results and the definition of obstruction is controversial in different centers. Furthermore, the indications of diuretic renography in the follow-up of hydronephrosis are another subject of debate. Doppler ultrasound is a practical and safe modality which can use in differentiating nonobstructive hydronephrosis. This study shows that RI ratio and delta RI with a high specificity could differentiate nonobstructive hydronephrosis, and therefore it is a promising way to use especially in the follow-up of children with hydronephrosis. Further extensive studies are helpful for determining whether Doppler ultrasound can replace diuretic renography in the follow-up of children after pyeloplasty or not.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.