- Email: [email protected]
Congenital hydronephrosis: Correlation of fetal ultrasonographic findings with infant outcome
Congenital hydronephrosis: Correlation of fetal ultrasonographic findings with infant outcome Jane E. Corteville, MD, Diana L. Gray, MD, and James P. Crane, MD St. Louis, Missouri
Although congenital hydronephrosis is a common fetal disorder, ultrasonographic criteria for prenatal diagnosis remain poorly defined. In this study prenatal ultrasonographic findings were correlated with postnatal outcome in 63 fetuses with suspected hydronephrosis. Prenatal ultrasonographic measurements included length, anteroposterior diameter, and transverse diameter of the kidney and renal pelvis, as well as dorsal renal parenchymal thickness. In 45 of the 63 fetuses, hydronephrosis was confirmed postnatally. These infants were divided into two groups on the basis of renal status: (1) abnormal renal function and/or surgery required (n = 31) and (2) normal renal function with no surgery required (n = 14). The anteroposterior diameter of the renal pelvis was the simplest and most sensitive technique for prenatal diagnosis of congenital hydronephrosis, allowing identification of 100% of cases. Postnatal follow-up studies are warranted if an anteroposterior pelvic diameter is ~4 mm before 33 weeks or ~ 7 mm after 33 weeks. (AM J OSSTET GVNECOL 1991 ;165:384-8.)
Key words: Congenital hydronephrosis, prenatal diagnosis, ultrasonography, infant outcome Congenital hydronephrosis is a commonly encountered birth defect, occurring with a frequency of 1 in every 330 infants (unpublished data). Clinical presentation is variable and may include a palpable abdominal mass, intermittent pain, vomiting, failure to thrive, or urinary tract infection. Surgical correction of congenital hydronephrosis within 6 to 12 months of birth is typically associated with good outcome whereas delayed diagnosis carries an increased risk of renal impairment. I, 2 Unfortunately, diagnosis on clinical grounds alone is often delayed with <25% of cases detected by age 1 and only 55% by age 5.' Prenatal ultrasonography affords the opportunity for early recognition and treatment of congenital hydronephrosis with hopeful preservation of renal function. While several ultrasonographic criteria have been suggested for prenatal diagnosis of congenital hydronephrosis!-8 few are based on long-term infant outcome and assessment of renal function. The purpose of this investigation was to assess the relative accuracy of these ultrasonographic techniques in the same population to From the Genetics Division, Department of Obstetrics and Gynecology, Washington University School of Medicine, Supported in part by a grant from the Jewish Hospital Endowment Fund. Presented at the Thirty-fourth Annual Convention of the American Institute of Ultrasound in Medicine, New Orleans, Louisiana, March 4-7,1990. Received for publication August 1 , 1990; revised December 26, 1990; accepted February 1,1991, Reprint requests: James p, Crane, MD, Genetics Division, Department of Obstetrics and Gynecology, Washington University School of Medicine, The Jewish Hospital of St. Louis, 216 So, Kingshighway Blvd., St. Louis, MO 63110.
611 /28481
384
Table I. Congenital hydronephrosis (N = 45)
I
Associated malformations Heart defects Pulmonary sequestration Pyloric stenosis Hypospadias
No. 6 (9.5%)
3 1 1 1
Trisomy 21
Table II. Congenital hydronephrosis: Site of obstruction Level of obstruction
Ureteropelvic junction Ureterovesical junction Ureterocele Other lesions*
Frequency 67% 14% 3% 14%
(40/60) (9160) (2/60) (9/60)
Ultrasonographic accuracy 97% (39/40)
33% (3/9) 50% (112) 56% (5/9)
*Isolated reflux, dilated extrarenal pelvis, renal duplication, multicystic renal dysplasia.
determine the best predictor(s) of renal compromise and/or the need for surgical intervention after birth.
Material and methods The study population was composed of 97 kidneys from 63 fetuses with suspected hydronephrosis on the basis of gross ultrasonographic appearance. All ultrasonographic examinations were conducted in the Washington University Obstetrical Ultrasound Laboratory between 1982 and 1988. Fetuses with bladder outlet obstruction were excluded. A total of 246 ultrasono-
Volume 165 Number 2
Congenital hydronephrosis-Ultrasonographic diagnosis 385
40r------------------------------4
E
.s
30
a: w
f--
W ~
«
is
20
()
:> ....J w
0...
0...
«
:~
••••••••
Fig. 1. Technique for ultrasonographic measurement of renal dimensions. A, Anteroposterior diameter of kidney; B, transverse diameter of kidney; C, dorsal parenchymal thickness; D, anteroposterior diameter of renal pelvis; E, transverse diameter of renal pelvis; F, length of renal pelvis; G, length of kidney.
graphic studies were performed for an average of 3.9 per fetus. The examinations were divided into three gestational age time frames (14 to 23.9, 24 to 32.9, and 33 to 42 weeks) with only a single examination per fetus included in each. The following renal measurements were obtained at each ultrasonographic examination: (1) kidney length, transverse diameter, and anteroposterior diameter; (2) renal pelvis length, transverse diameter, and anteroposterior diameter; (3) maximal renal parenchymal thickness. Kidney and renal pelvis lengths were measured in a coronal plane whereas transverse and anteroposterior dimensions were measured in an axial plane through the fetal abdomen. Renal parenchymal thickness was measured along the dorsum of the kidney when viewed in an axial plane (Fig. 1). Infant outcome was collected by chart review. The mean duration of follow-up was 20 months with a range of 1 to 71 months. Primary end points included: (1) the accuracy of each ultrasonographic parameter in diagnosing congenital hydronephrosis and (2) the ability of ultrasonography to predict adverse outcome as determined by the presence of renal compromise and/or need for surgical intervention. Postnatal con-
.
.
0~~~~~~~=L~±=6J 14
18
22
26
30
34
38
42
GESTATIONAL AGE (Weeks) Fig. 2. Natural history of congenital hydronephrosis. Case 39 (e) was classified as having progessive lesion. Fetus in case 32 (A.) had stable hydronephrosis with serial anteroposterior pelvic diameters that paralleled normal reference range. In case 55 (.) spontaneous resolution was observed.
genital hydronephrosis was confirmed by ultrasonography, intravenous pyelography, or renal scintigraphy. Criteria for defining renal compromise included any of the following findings: (1) serum creatinine level >0.7 mg/dl more than 7 days after delivery; (2) cortical atrophy identified by intravenous pyelography, postnatal sonogram, or renal scintillation scan; (3) differential excretion >20% on renal scintillation scan, or (4) significant urinary tract infection (more than two in female infants and more than one in male infants). To evaluate which fetal renal dimensions best predict outcome, the following ultrasonographic parameters were examined: (1) anteroposterior diameter of the renal pelvis, (2) ratio of anteroposterior diameters of renal pelvis to kidney, (3) ratio of transverse diameters of the renal pelvis to kidney, (4) kidney circumference/abdominal circumference ratio, and (5) maximal renal parenchymal thickness. Multiple thresholds were analyzed for each of these parameters and correlated with infant outcome. The goal was to choose the threshold with the highest sensitivity for predicting congenital hydronephrosis to minimize the chance of missing a clinically significant lesion. False-positive rates for each threshold also were determined. Secondary analyses included the clinical significance of prenatally detected calicectasis, the accuracy of pre-
386
Corteville, Gray, and Crane
August 1991 Am J Obstet Gynecol
Table III. Anteroposterior pelvic diameter versus risk of postnatal congenital hydronephrosis 14-23 wk A nteroposterior pelvic diameter (mm) ~4
~7 ~1O
Sensitivity (%)
I
24-32 wk
False-positive rate (%)
Sensitivity
55 35
100 61
32
18
33-42 wk
False-positive rate (%)
Sensitivity
100
42
74
14
100 100 82
(%)
I
34
91
(%)
I
False-positive rate (%)
24
21 18
Table IV. Anteroposterior diameter pelvis/kidney versus risk of postnatal congenital hydronephrosis 14-22 wk A nteroposterior diameter, pelvis / kidney ratio
Sensitivity
>0.28 >0.30 >0.40 >0.50
100
(%)
92
64
40
I
23-32 wk
False-positive rate (%)
Sensitivity
50
90 88 69
48
47 47
natal ultrasonography in predicting the level of obstruction, the frequency of associated malformation, and the natural history of the obstructive process during fetal life. Results
The data base consisted of97 kidneys from 63 fetuses with suspected congenital hydronephrosis on the basis of gross ultrasonographic appearance. In 56% of cases this was an incidental finding during obstetric ultrasonographic examination for other indications. The remaining 44% of fetuses were referred because of a suspected renal anomaly identified during a prior outside ultrasonographic examination. Congenital hydronephrosis was confirmed by postnatal evaluation in 45 (71 %) of the 63 fetuses. Bilateral lesions were present in IS (33%) of the 45 infants, for a total of 60 index kidneys. Among unilateral cases the left kidney was more commonly involved (57%). As expected, a strong male predilection was found with a male/female ratio of 4.5: 1. A single umbilical arery was present in 3.2% of cases. Major nonrenal malformations were identified in 9.5% of infants with congenital heart disease most commonly observed (Table I). Ureteropelvic junction obstruction was the most common underlying lesion, accounting for nearly two thirds of all cases (Table II). The ureterovesical junction was the next most common site of obstruction (14%), followed by ureterocele (5%) and a variety of other lesions. The accuracy of fetal ultrasonography in predicting the level of obstruction varied widely by underlying lesion. For example, 97% of ureteropelvic junction obstructions were correctly identified versus only 33% of ureterovesical junction obstructions (Table II). The natural history of congenital hydronephrosis
(%)
52
I
33-42 wk
False-positive rate (%)
Sensitivity
37 33 15 8
91 84 68
(%)
50
I
False-positive rate (%)
20 20 21 15
was established by analyzing data from 39 hydronephrotic kidneys monitored with serial ultrasonographic studies. A graph was established for each kidney; the anteroposterior renal pelvic diameter was plotted against gestational age and compared with a similar nomogram derived from a control population of 137 normal fetuses. Congenital hydronephrosis was considered progressive if the observed slope exceeded that of the reference curve (Fig. 2). With this approach, 14 of 39 obstructive lesions (36%) were considered stable throughout gestation whereas progressive obstruction was noted in 37% (14/39) and some improvement in the remaining 28% (11/39). Amniotic fluid volume was normal in 87% (39/45) of hydronephrotic pregnancies whereas oligohydramnios was noted in four pregnancies (9%) and polyhydramnios in two cases (4%). Six different ultrasonographic parameters were evaluated to determine the best predictor(s) of postnatally confirmed congenital hydronephrosis and related morbidity. Anteroposterior renal pelvic diameter. In analyzing the data three different thresholds were examined: 2:4, 2:7, and 2: 10 mm (Table III). Before 33 weeks' gestation, a threshold of 2:4 mm was required to achieve a sensitivity of 100%. During the last 7 weeks of pregnancy, a threshold of 2:7 mm could be used while still maintaining a sensitivity of 100%. Anteroposterior diameter of renal pelvis/kidney ratio. Thresholds of >0.28, >0.30, >0.40, and >0.50 were analyzed (Table IV). An anteroposterior pelvis/kidney ratio of >0.28 was required to achieve a sensitivity of 90% to 100%. Transverse diameter of renal pelvis/kidney ratio. Thresholds of >0.3, >0.4, and >0.5 were selected for analysis (Table V). The threshold of >0.3 proved most sensitive, identifying 96% to 100% of fetuses with postnatally confirmed cong-enital hydroneDhrosis.
Congenital hydronephrosis-Ultrasonographic diagnosis
Volume 165 Number 2
387
Table V. Transverse diameter of pelvis/kidney versus risk of postnatal congenital hydronephrosis 14-23 wk Transverse diameter, pelvis / kidney ratio
Sensitivity
>0.30 >0.40 >0.50
96 72 52
(%)
I
24-32 wk
False-positive rate (%)
Sensitivity
51 51
100 86 60
41
I
(%)
33-42 wk
False-positive rate (%)
Sensitivity
37
100 88 77
35 31
(%)
1
False-positive rate (%)
26
27
23
Table VI. Kidney circumference / abdominal circumference versus risk of postnatal congenital hydronephrosis 14-23 wk
Kidney circumference / abdominal circumference ratio
Sensitivity
>0.28 >0.30 >0.32
90
24-32 wk
False-positive rate (%)
Sensitivity
95
49
89
76
43
58
(%)
I
44
Kidney / abdominal circumference ratio. Thresholds of>0.28, >0.30, and >0.32 were examined (Table VI). A threshold of >0.28 appeared optimal although sensitivity decreased to only 59% in late pregnancy. Renal parenchymal thickness. Although ultrasonographically measured parenchymal thickness might potentially reflect cortical atrophy, no threshold that accurately predicted infant outcome could be identified. Calicectasis. Dilated calyces were noted prenatally in only 47% (28/60) of hydronephrotic kidneys. While not highly sensitive for prenatal diagnosis of postnatally confirmed congenital hydronephrosis, calicectasis did correlate with morbidity. As noted in Table VII, surgery and renal compromise were significantly more common in fetuses with dilated calyces than in fetuses without calicectasis.
Comment Ultrasonographic recognition of a prominent fluid collection within the fetal renal pelvis represents a diagnostic dilemma. False-positive diagnosis of fetal hydronephrosis may lead to unwarranted parental anxiety, as well as unnecessary and expensive diagnostic studies after delivery. At the same time failure to recognize and promptly treat clinically significant obstructions may result in impaired renal function. Several ultrasonographic criteria have been proposed for prenatal diagnosis of congenital hydronephrosis; however, this is the first investigation to compare these parameters in the same population and correlate their accuracy in predicting infant outcome. Anteroposterior renal pelvic diameter is perhaps the most commonly used technique for diagnosing congenital hydronephrosis. Both Arger et al. 4 and Grignon et al. 5 considered an anteroposterior diameter ~ 10 mm to be predictive of abnormal outcome. In addition, Ar-
(%)
1
33-42 wk
False-positive rate (%)
Sensitivity
33
59
27
41
32
76
(%)
48
1
False-positive rate (%)
26 22
o
Table VII. Impact of calicectasis on postnatal outcome
Kidneys (No.) Surgery required (%)
Renal compromise (%)
Surgery with or without compromise (%)
Normal calyces
Dilated calyces
32
28
28
50
56
89
41
79
ger et al. i suggested follow-up scans in any fetus with an anteroposterior diameter of 25 mm. Data from the current investigation indicate that a threshold of 10 mm is too high. For example, 68% of fetuses with postnatally confirmed congenital hydronephrosis would be missed before 24 weeks' gestation (Table III). In contrast, a threshold of 4 mm yields a sensitivity of 100%. While the false-positive rate is higher with a 4 mm threshold, the goal is to identify all "at risk" fetuses. In addition, the false-positive rate declines as gestational age advances, falling to only 24% near term. After 33 weeks' gestation, a threshold of 7 mm can be used while still maintaining 100% sensitivity. The second analyzed ultrasonographic parameter was the ratio of the anteroposterior diameters of the renal pelvis and kidney. Both Arger et al. i and Kleiner et al. 6 suggested that a ratio exceeding 0.50 should be used for prenatal diagnosis of congenital hydronephrosis. Before 24 weeks' gestation, this threshold would fail to detect 60% of postnatally confirmed lesions; after 24 weeks, nearly 50% of affected fetuses would remain undetected (Table IV). To achieve a sensitivity of 90% to 100%, an anteroposterior pelvis/kid-
388
Corteville, Gray, and Crane
August 1991 Am J Obstet Gynecol
Table VIII. Anteroposterior pelvic diameter: Thresholds for counseling 14-23 wk
Anteroposterior pelvic diameter (mm) <3
4-6 7-9
>10
CH (%)
o
41 53 82
I
24-32 wk
SIC
CH
(%)
(%)
o
o
19 40 73
I
33-42 wk
SIC
CH
(%)
(%)
o
o
6
67 82
38
13
86
72
33
o
I
SIC (%)
o
o
50 59
CH, Congenital hydronephrosis confirmed postnatally; SIC, postnatal surgery andlor evidence of renal compromise.
ney ratio of 0.28 is recommended. While this ratio is associated with a high false-positive rate in early pregnancy, progressive decline in the rate of false-positive results is noted as gestation advances. The third examined biometric approach to prenatal diagnosis of congenital hydronephrosis was the ratio of the transverse diameters of the renal pelvis versus the kidney. Jeanty and Romer0 7 described this technique in 1984 and considered a ratio 2:0.5 to be abnormal. In our study population this threshold identified only 52% of confirmed lesions before 24 weeks' gestation and missed 23% to 40% of affected fetuses during later pregnancy (Table V). A transverse pelvis/kidney ratio of 0.30 appeared more ideal, identifying nearly 100% of cases at all gestational ages. As with previously cited techniques, the false-positive rate declines with advancing gestational age. The kidney circumference/abdominal circumference ratio was originally described by Grannum et aJ.8 Although this method was developed as a means of quantifying renal enlargement in cystic kidney disease, an increased ratio also might be anticipated in congenital hydronephrosis. On examination of our data, a kidney circumference/abdominal circumference ratio of 0.28 appeared optimal during the midtrimester although sensitivity decreased to only 59% during late pregnancy (Table VI). Of all analyzed techniques for prenatal diagnosis of congenital hydronephrosis, the anteroposterior diameter of the renal pelvis is simplest and has sensitivity and false-positive rates that are equal or superior to those of other ultrasonographic approaches. Follow-up studies are warranted in 3 to 4 weeks if an anteroposterior pelvic diameter is 2:4 mm before 33 weeks or 2:7 mm after 33 weeks. Postnatal follow-up studies are recommended if persistent renal pelvic dilatation is evident. Neonatal ultrasonographic evaluation should be postponed for several days after delivery to assure that the infant is well hydrated and the full extent of any potential obstruction appreciated. If abnormal, additional diagnostic studies including intravenous pyelography and renal scintigraphy may be helpful in further
defining the lesion, as well as in delineating appropriate medical management. Table VIII provides useful information for counseling parents when a diagnosis of congenital hydronephrosis is suspected. For example, before 24 weeks' gestation, there is a 41 % chance of postnatally confirmed congenital hydronephrosis if the anteroposterior pelvic diameter is between 4 and 6 mm and a 19% risk for impaired renal function and/or surgery. The risk for adverse outcome increases to 40% if the anteroposterior diameter is between 7 and 9 mm and approaches 73% if 2:10 mm. It is clear from this study that congenital hydronephrosis is a clinically significant birth defect. Overall, 69% (31/45) of affected infants had related morbidity as defined by abnormal renal function and/or the need for surgical intervention. We hope this investigation clarifies appropriate ultrasonographic criteria for diagnosis of congenital hydronephrosis and provides useful guidelines for counseling and management of this common congenital malformation. REFERENCES 1. Mayor G, Genton N, Torrado A, GuignardJP. Renal func-
2. 3. 4. 5. 6. 7. 8.
tion in obstructive nephropathy: long-term effect of reconstructive surgery. Pediatrics 1975;56:740-7. Mathieu H, Loirat C, Macher MA, et al. Long-term outcome of children with malformative uropathies. lnt] Pediatr Nephrol 1985;6:3-12. Snyder HM, Lebowitz RL, Colodny AH, Bauer SB, Retik AB. Ureteropelvic junction obstruction in children. Urol Clin North Am 1980;7:273-89. Arger PH, Coleman BG, Mintz MC, et al. Routine fetal genitourinary tract screening. Radiology 1985; 156:485-9. Grignon A, Filion R, Filiatrault D, et al. Urinary tract dilatation in utero: classification and clinical applications. Radiology 1986;160:645-7. Kleiner B, Callen PW, Filly RA. Sonographic analysis of the fetus with ureteropelvic junction obstruction. AJR 1987;148:359-62. Jeanty P, Romero R. Are the kidneys normal? In: Brehm J], Boynton S, eds. Obstetrical ultrasound. New York: McGraw-Hill,1984:143-9. Grannum P, Bracken M, Silverman R, Hobbins ]C. Assessment of fetal kidney size in normal gestation by comparison of ratio of kidney circumference to abdominal circumference. AM] OBSTET GYNECOL 1980;136:249-54.
Although congenital hydronephrosis is a common fetal disorder, ultrasonographic criteria for prenatal diagnosis remain poorly defined. In this study prenatal ultrasonographic findings were correlated with postnatal outcome in 63 fetuses with suspected hydronephrosis. Prenatal ultrasonographic measurements included length, anteroposterior diameter, and transverse diameter of the kidney and renal pelvis, as well as dorsal renal parenchymal thickness. In 45 of the 63 fetuses, hydronephrosis was confirmed postnatally. These infants were divided into two groups on the basis of renal status: (1) abnormal renal function and/or surgery required (n = 31) and (2) normal renal function with no surgery required (n = 14). The anteroposterior diameter of the renal pelvis was the simplest and most sensitive technique for prenatal diagnosis of congenital hydronephrosis, allowing identification of 100% of cases. Postnatal follow-up studies are warranted if an anteroposterior pelvic diameter is ~4 mm before 33 weeks or ~ 7 mm after 33 weeks. (AM J OSSTET GVNECOL 1991 ;165:384-8.)
Key words: Congenital hydronephrosis, prenatal diagnosis, ultrasonography, infant outcome Congenital hydronephrosis is a commonly encountered birth defect, occurring with a frequency of 1 in every 330 infants (unpublished data). Clinical presentation is variable and may include a palpable abdominal mass, intermittent pain, vomiting, failure to thrive, or urinary tract infection. Surgical correction of congenital hydronephrosis within 6 to 12 months of birth is typically associated with good outcome whereas delayed diagnosis carries an increased risk of renal impairment. I, 2 Unfortunately, diagnosis on clinical grounds alone is often delayed with <25% of cases detected by age 1 and only 55% by age 5.' Prenatal ultrasonography affords the opportunity for early recognition and treatment of congenital hydronephrosis with hopeful preservation of renal function. While several ultrasonographic criteria have been suggested for prenatal diagnosis of congenital hydronephrosis!-8 few are based on long-term infant outcome and assessment of renal function. The purpose of this investigation was to assess the relative accuracy of these ultrasonographic techniques in the same population to From the Genetics Division, Department of Obstetrics and Gynecology, Washington University School of Medicine, Supported in part by a grant from the Jewish Hospital Endowment Fund. Presented at the Thirty-fourth Annual Convention of the American Institute of Ultrasound in Medicine, New Orleans, Louisiana, March 4-7,1990. Received for publication August 1 , 1990; revised December 26, 1990; accepted February 1,1991, Reprint requests: James p, Crane, MD, Genetics Division, Department of Obstetrics and Gynecology, Washington University School of Medicine, The Jewish Hospital of St. Louis, 216 So, Kingshighway Blvd., St. Louis, MO 63110.
611 /28481
384
Table I. Congenital hydronephrosis (N = 45)
I
Associated malformations Heart defects Pulmonary sequestration Pyloric stenosis Hypospadias
No. 6 (9.5%)
3 1 1 1
Trisomy 21
Table II. Congenital hydronephrosis: Site of obstruction Level of obstruction
Ureteropelvic junction Ureterovesical junction Ureterocele Other lesions*
Frequency 67% 14% 3% 14%
(40/60) (9160) (2/60) (9/60)
Ultrasonographic accuracy 97% (39/40)
33% (3/9) 50% (112) 56% (5/9)
*Isolated reflux, dilated extrarenal pelvis, renal duplication, multicystic renal dysplasia.
determine the best predictor(s) of renal compromise and/or the need for surgical intervention after birth.
Material and methods The study population was composed of 97 kidneys from 63 fetuses with suspected hydronephrosis on the basis of gross ultrasonographic appearance. All ultrasonographic examinations were conducted in the Washington University Obstetrical Ultrasound Laboratory between 1982 and 1988. Fetuses with bladder outlet obstruction were excluded. A total of 246 ultrasono-
Volume 165 Number 2
Congenital hydronephrosis-Ultrasonographic diagnosis 385
40r------------------------------4
E
.s
30
a: w
f--
W ~
«
is
20
()
:> ....J w
0...
0...
«
:~
••••••••
Fig. 1. Technique for ultrasonographic measurement of renal dimensions. A, Anteroposterior diameter of kidney; B, transverse diameter of kidney; C, dorsal parenchymal thickness; D, anteroposterior diameter of renal pelvis; E, transverse diameter of renal pelvis; F, length of renal pelvis; G, length of kidney.
graphic studies were performed for an average of 3.9 per fetus. The examinations were divided into three gestational age time frames (14 to 23.9, 24 to 32.9, and 33 to 42 weeks) with only a single examination per fetus included in each. The following renal measurements were obtained at each ultrasonographic examination: (1) kidney length, transverse diameter, and anteroposterior diameter; (2) renal pelvis length, transverse diameter, and anteroposterior diameter; (3) maximal renal parenchymal thickness. Kidney and renal pelvis lengths were measured in a coronal plane whereas transverse and anteroposterior dimensions were measured in an axial plane through the fetal abdomen. Renal parenchymal thickness was measured along the dorsum of the kidney when viewed in an axial plane (Fig. 1). Infant outcome was collected by chart review. The mean duration of follow-up was 20 months with a range of 1 to 71 months. Primary end points included: (1) the accuracy of each ultrasonographic parameter in diagnosing congenital hydronephrosis and (2) the ability of ultrasonography to predict adverse outcome as determined by the presence of renal compromise and/or need for surgical intervention. Postnatal con-
.
.
0~~~~~~~=L~±=6J 14
18
22
26
30
34
38
42
GESTATIONAL AGE (Weeks) Fig. 2. Natural history of congenital hydronephrosis. Case 39 (e) was classified as having progessive lesion. Fetus in case 32 (A.) had stable hydronephrosis with serial anteroposterior pelvic diameters that paralleled normal reference range. In case 55 (.) spontaneous resolution was observed.
genital hydronephrosis was confirmed by ultrasonography, intravenous pyelography, or renal scintigraphy. Criteria for defining renal compromise included any of the following findings: (1) serum creatinine level >0.7 mg/dl more than 7 days after delivery; (2) cortical atrophy identified by intravenous pyelography, postnatal sonogram, or renal scintillation scan; (3) differential excretion >20% on renal scintillation scan, or (4) significant urinary tract infection (more than two in female infants and more than one in male infants). To evaluate which fetal renal dimensions best predict outcome, the following ultrasonographic parameters were examined: (1) anteroposterior diameter of the renal pelvis, (2) ratio of anteroposterior diameters of renal pelvis to kidney, (3) ratio of transverse diameters of the renal pelvis to kidney, (4) kidney circumference/abdominal circumference ratio, and (5) maximal renal parenchymal thickness. Multiple thresholds were analyzed for each of these parameters and correlated with infant outcome. The goal was to choose the threshold with the highest sensitivity for predicting congenital hydronephrosis to minimize the chance of missing a clinically significant lesion. False-positive rates for each threshold also were determined. Secondary analyses included the clinical significance of prenatally detected calicectasis, the accuracy of pre-
386
Corteville, Gray, and Crane
August 1991 Am J Obstet Gynecol
Table III. Anteroposterior pelvic diameter versus risk of postnatal congenital hydronephrosis 14-23 wk A nteroposterior pelvic diameter (mm) ~4
~7 ~1O
Sensitivity (%)
I
24-32 wk
False-positive rate (%)
Sensitivity
55 35
100 61
32
18
33-42 wk
False-positive rate (%)
Sensitivity
100
42
74
14
100 100 82
(%)
I
34
91
(%)
I
False-positive rate (%)
24
21 18
Table IV. Anteroposterior diameter pelvis/kidney versus risk of postnatal congenital hydronephrosis 14-22 wk A nteroposterior diameter, pelvis / kidney ratio
Sensitivity
>0.28 >0.30 >0.40 >0.50
100
(%)
92
64
40
I
23-32 wk
False-positive rate (%)
Sensitivity
50
90 88 69
48
47 47
natal ultrasonography in predicting the level of obstruction, the frequency of associated malformation, and the natural history of the obstructive process during fetal life. Results
The data base consisted of97 kidneys from 63 fetuses with suspected congenital hydronephrosis on the basis of gross ultrasonographic appearance. In 56% of cases this was an incidental finding during obstetric ultrasonographic examination for other indications. The remaining 44% of fetuses were referred because of a suspected renal anomaly identified during a prior outside ultrasonographic examination. Congenital hydronephrosis was confirmed by postnatal evaluation in 45 (71 %) of the 63 fetuses. Bilateral lesions were present in IS (33%) of the 45 infants, for a total of 60 index kidneys. Among unilateral cases the left kidney was more commonly involved (57%). As expected, a strong male predilection was found with a male/female ratio of 4.5: 1. A single umbilical arery was present in 3.2% of cases. Major nonrenal malformations were identified in 9.5% of infants with congenital heart disease most commonly observed (Table I). Ureteropelvic junction obstruction was the most common underlying lesion, accounting for nearly two thirds of all cases (Table II). The ureterovesical junction was the next most common site of obstruction (14%), followed by ureterocele (5%) and a variety of other lesions. The accuracy of fetal ultrasonography in predicting the level of obstruction varied widely by underlying lesion. For example, 97% of ureteropelvic junction obstructions were correctly identified versus only 33% of ureterovesical junction obstructions (Table II). The natural history of congenital hydronephrosis
(%)
52
I
33-42 wk
False-positive rate (%)
Sensitivity
37 33 15 8
91 84 68
(%)
50
I
False-positive rate (%)
20 20 21 15
was established by analyzing data from 39 hydronephrotic kidneys monitored with serial ultrasonographic studies. A graph was established for each kidney; the anteroposterior renal pelvic diameter was plotted against gestational age and compared with a similar nomogram derived from a control population of 137 normal fetuses. Congenital hydronephrosis was considered progressive if the observed slope exceeded that of the reference curve (Fig. 2). With this approach, 14 of 39 obstructive lesions (36%) were considered stable throughout gestation whereas progressive obstruction was noted in 37% (14/39) and some improvement in the remaining 28% (11/39). Amniotic fluid volume was normal in 87% (39/45) of hydronephrotic pregnancies whereas oligohydramnios was noted in four pregnancies (9%) and polyhydramnios in two cases (4%). Six different ultrasonographic parameters were evaluated to determine the best predictor(s) of postnatally confirmed congenital hydronephrosis and related morbidity. Anteroposterior renal pelvic diameter. In analyzing the data three different thresholds were examined: 2:4, 2:7, and 2: 10 mm (Table III). Before 33 weeks' gestation, a threshold of 2:4 mm was required to achieve a sensitivity of 100%. During the last 7 weeks of pregnancy, a threshold of 2:7 mm could be used while still maintaining a sensitivity of 100%. Anteroposterior diameter of renal pelvis/kidney ratio. Thresholds of >0.28, >0.30, >0.40, and >0.50 were analyzed (Table IV). An anteroposterior pelvis/kidney ratio of >0.28 was required to achieve a sensitivity of 90% to 100%. Transverse diameter of renal pelvis/kidney ratio. Thresholds of >0.3, >0.4, and >0.5 were selected for analysis (Table V). The threshold of >0.3 proved most sensitive, identifying 96% to 100% of fetuses with postnatally confirmed cong-enital hydroneDhrosis.
Congenital hydronephrosis-Ultrasonographic diagnosis
Volume 165 Number 2
387
Table V. Transverse diameter of pelvis/kidney versus risk of postnatal congenital hydronephrosis 14-23 wk Transverse diameter, pelvis / kidney ratio
Sensitivity
>0.30 >0.40 >0.50
96 72 52
(%)
I
24-32 wk
False-positive rate (%)
Sensitivity
51 51
100 86 60
41
I
(%)
33-42 wk
False-positive rate (%)
Sensitivity
37
100 88 77
35 31
(%)
1
False-positive rate (%)
26
27
23
Table VI. Kidney circumference / abdominal circumference versus risk of postnatal congenital hydronephrosis 14-23 wk
Kidney circumference / abdominal circumference ratio
Sensitivity
>0.28 >0.30 >0.32
90
24-32 wk
False-positive rate (%)
Sensitivity
95
49
89
76
43
58
(%)
I
44
Kidney / abdominal circumference ratio. Thresholds of>0.28, >0.30, and >0.32 were examined (Table VI). A threshold of >0.28 appeared optimal although sensitivity decreased to only 59% in late pregnancy. Renal parenchymal thickness. Although ultrasonographically measured parenchymal thickness might potentially reflect cortical atrophy, no threshold that accurately predicted infant outcome could be identified. Calicectasis. Dilated calyces were noted prenatally in only 47% (28/60) of hydronephrotic kidneys. While not highly sensitive for prenatal diagnosis of postnatally confirmed congenital hydronephrosis, calicectasis did correlate with morbidity. As noted in Table VII, surgery and renal compromise were significantly more common in fetuses with dilated calyces than in fetuses without calicectasis.
Comment Ultrasonographic recognition of a prominent fluid collection within the fetal renal pelvis represents a diagnostic dilemma. False-positive diagnosis of fetal hydronephrosis may lead to unwarranted parental anxiety, as well as unnecessary and expensive diagnostic studies after delivery. At the same time failure to recognize and promptly treat clinically significant obstructions may result in impaired renal function. Several ultrasonographic criteria have been proposed for prenatal diagnosis of congenital hydronephrosis; however, this is the first investigation to compare these parameters in the same population and correlate their accuracy in predicting infant outcome. Anteroposterior renal pelvic diameter is perhaps the most commonly used technique for diagnosing congenital hydronephrosis. Both Arger et al. 4 and Grignon et al. 5 considered an anteroposterior diameter ~ 10 mm to be predictive of abnormal outcome. In addition, Ar-
(%)
1
33-42 wk
False-positive rate (%)
Sensitivity
33
59
27
41
32
76
(%)
48
1
False-positive rate (%)
26 22
o
Table VII. Impact of calicectasis on postnatal outcome
Kidneys (No.) Surgery required (%)
Renal compromise (%)
Surgery with or without compromise (%)
Normal calyces
Dilated calyces
32
28
28
50
56
89
41
79
ger et al. i suggested follow-up scans in any fetus with an anteroposterior diameter of 25 mm. Data from the current investigation indicate that a threshold of 10 mm is too high. For example, 68% of fetuses with postnatally confirmed congenital hydronephrosis would be missed before 24 weeks' gestation (Table III). In contrast, a threshold of 4 mm yields a sensitivity of 100%. While the false-positive rate is higher with a 4 mm threshold, the goal is to identify all "at risk" fetuses. In addition, the false-positive rate declines as gestational age advances, falling to only 24% near term. After 33 weeks' gestation, a threshold of 7 mm can be used while still maintaining 100% sensitivity. The second analyzed ultrasonographic parameter was the ratio of the anteroposterior diameters of the renal pelvis and kidney. Both Arger et al. i and Kleiner et al. 6 suggested that a ratio exceeding 0.50 should be used for prenatal diagnosis of congenital hydronephrosis. Before 24 weeks' gestation, this threshold would fail to detect 60% of postnatally confirmed lesions; after 24 weeks, nearly 50% of affected fetuses would remain undetected (Table IV). To achieve a sensitivity of 90% to 100%, an anteroposterior pelvis/kid-
388
Corteville, Gray, and Crane
August 1991 Am J Obstet Gynecol
Table VIII. Anteroposterior pelvic diameter: Thresholds for counseling 14-23 wk
Anteroposterior pelvic diameter (mm) <3
4-6 7-9
>10
CH (%)
o
41 53 82
I
24-32 wk
SIC
CH
(%)
(%)
o
o
19 40 73
I
33-42 wk
SIC
CH
(%)
(%)
o
o
6
67 82
38
13
86
72
33
o
I
SIC (%)
o
o
50 59
CH, Congenital hydronephrosis confirmed postnatally; SIC, postnatal surgery andlor evidence of renal compromise.
ney ratio of 0.28 is recommended. While this ratio is associated with a high false-positive rate in early pregnancy, progressive decline in the rate of false-positive results is noted as gestation advances. The third examined biometric approach to prenatal diagnosis of congenital hydronephrosis was the ratio of the transverse diameters of the renal pelvis versus the kidney. Jeanty and Romer0 7 described this technique in 1984 and considered a ratio 2:0.5 to be abnormal. In our study population this threshold identified only 52% of confirmed lesions before 24 weeks' gestation and missed 23% to 40% of affected fetuses during later pregnancy (Table V). A transverse pelvis/kidney ratio of 0.30 appeared more ideal, identifying nearly 100% of cases at all gestational ages. As with previously cited techniques, the false-positive rate declines with advancing gestational age. The kidney circumference/abdominal circumference ratio was originally described by Grannum et aJ.8 Although this method was developed as a means of quantifying renal enlargement in cystic kidney disease, an increased ratio also might be anticipated in congenital hydronephrosis. On examination of our data, a kidney circumference/abdominal circumference ratio of 0.28 appeared optimal during the midtrimester although sensitivity decreased to only 59% during late pregnancy (Table VI). Of all analyzed techniques for prenatal diagnosis of congenital hydronephrosis, the anteroposterior diameter of the renal pelvis is simplest and has sensitivity and false-positive rates that are equal or superior to those of other ultrasonographic approaches. Follow-up studies are warranted in 3 to 4 weeks if an anteroposterior pelvic diameter is 2:4 mm before 33 weeks or 2:7 mm after 33 weeks. Postnatal follow-up studies are recommended if persistent renal pelvic dilatation is evident. Neonatal ultrasonographic evaluation should be postponed for several days after delivery to assure that the infant is well hydrated and the full extent of any potential obstruction appreciated. If abnormal, additional diagnostic studies including intravenous pyelography and renal scintigraphy may be helpful in further
defining the lesion, as well as in delineating appropriate medical management. Table VIII provides useful information for counseling parents when a diagnosis of congenital hydronephrosis is suspected. For example, before 24 weeks' gestation, there is a 41 % chance of postnatally confirmed congenital hydronephrosis if the anteroposterior pelvic diameter is between 4 and 6 mm and a 19% risk for impaired renal function and/or surgery. The risk for adverse outcome increases to 40% if the anteroposterior diameter is between 7 and 9 mm and approaches 73% if 2:10 mm. It is clear from this study that congenital hydronephrosis is a clinically significant birth defect. Overall, 69% (31/45) of affected infants had related morbidity as defined by abnormal renal function and/or the need for surgical intervention. We hope this investigation clarifies appropriate ultrasonographic criteria for diagnosis of congenital hydronephrosis and provides useful guidelines for counseling and management of this common congenital malformation. REFERENCES 1. Mayor G, Genton N, Torrado A, GuignardJP. Renal func-
2. 3. 4. 5. 6. 7. 8.
tion in obstructive nephropathy: long-term effect of reconstructive surgery. Pediatrics 1975;56:740-7. Mathieu H, Loirat C, Macher MA, et al. Long-term outcome of children with malformative uropathies. lnt] Pediatr Nephrol 1985;6:3-12. Snyder HM, Lebowitz RL, Colodny AH, Bauer SB, Retik AB. Ureteropelvic junction obstruction in children. Urol Clin North Am 1980;7:273-89. Arger PH, Coleman BG, Mintz MC, et al. Routine fetal genitourinary tract screening. Radiology 1985; 156:485-9. Grignon A, Filion R, Filiatrault D, et al. Urinary tract dilatation in utero: classification and clinical applications. Radiology 1986;160:645-7. Kleiner B, Callen PW, Filly RA. Sonographic analysis of the fetus with ureteropelvic junction obstruction. AJR 1987;148:359-62. Jeanty P, Romero R. Are the kidneys normal? In: Brehm J], Boynton S, eds. Obstetrical ultrasound. New York: McGraw-Hill,1984:143-9. Grannum P, Bracken M, Silverman R, Hobbins ]C. Assessment of fetal kidney size in normal gestation by comparison of ratio of kidney circumference to abdominal circumference. AM] OBSTET GYNECOL 1980;136:249-54.