Color and Power Doppler Sonography Versus DMSA Scintigraphy in Acute Pyelonephritis and in Prediction of Renal Scarring

MATERIALS AND METHODS

Clinical and Biologic Findings

From July 1997 to April 2000, we prospectively evaluated 57 children (7 boys, 50 girls; age range, 6 mo to 15.5 y; mean age ± SD, 5 ± 3 y; 1 child with a single kidney). This study was performed according to the Helsinki criteria and with the approval of the local Ethics Committee and the informed consent of both parents. To be included, a child had to have clinical findings consistent with APN, be older than 6 mo, have undergone DS at the acute stage of pyelonephritis (DS 1), and have undergone DMSA at the acute stage of pyelonephritis (DMSA 1). Each child received a standardized clinical examination including abdominal and lumbar fossa palpation, temperature and blood pressure measurement, and standardized blood studies including ionography, a differential blood count, a blood culture, and C-reactive protein (CRP) analysis. A urine sample (a clean catch in midstream) was taken for dipstick analysis and culturing. Usual intravenous antibiotic treatment was started immediately after clinical examination, and blood samples and cultures were obtained. APN was diagnosed if the patient had abdominal or lumbar fossa pain, fever > 38°C, and positive results from the urine culture (i.e., >10 white blood cells per cubic millimeter and bacteriuria to the extent of ≥10⁴ colony-forming units per milliliter). Children were excluded from the study if they had urinary tract obstruction, greater than grade III vesicoureteral reflux according to the International Grading Study (11), or breakthrough infection between inclusion and follow-up. Figure 1 summarizes the examination schedule.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Examination schedule. VCUG = voiding cystourethrography.

RESULTS

DISCUSSION

An early and reliable diagnostic tool for APN would be useful for selecting those children with urinary tract infection who require intravenous antibiotics. Such an examination and treatment could decrease the prevalence of renal scarring and its long-term complications. DMSA scintigraphy and CT have long been known to be highly sensitive and specific (2,3,7,15). Sonography is usually performed on an emergency basis to rule out pyohydronephrosis. Recently, gray-scale and DS findings associated with pyelonephritis have been described (8,16,17). However, the sensitivity of sonography was shown to remain inferior to that of DMSA, and sonography was therefore not considered a reference method (18). Roberts (19) described the “vascular phase” (slow vascular flow) of pyelonephritis, which was thought to represent the basis for the DS findings, which corresponded to the DMSA findings described by Majd and Rushton (20). In our prospective work, we compared the diagnostic values of DS and DMSA and particularly tested the following hypothesis: In children with APN and positive sonographic findings during the acute phase, subsequent scarring could be prone to develop in the same location.

Our population data confirmed the established predominance of girls (88%) among children (excluding neonates) with APN. Children who were <6 mo old were excluded because the pathophysiology is thought to be different in infants (21). Patients with known obstruction and high-grade reflux were not included because we wished to select children with pyelonephritis developing in a nondilated urinary tract. Biologic findings were consistent with pyelonephritis in the acute phase in all. The predictive values of body temperature and biologic tests were calculated. In no patient did any of these parameters provide pertinent information on the risk of scarring. Children with abnormal DMSA 1 findings had a significantly higher body temperature and a more severe biologic infection (Table 4) than did those with normal DMSA 1 findings. Only the CRP value correlated significantly with the DS 1 result. Such a difference may indicate that DMSA is more accurate for diagnosing APN. In spite of a 51% rate of scarring (abnormal DMSA 2 findings) among the children who were followed up, in none of these had any clinical symptom, such as hypertension or biologic deterioration of renal function, yet developed. Perhaps the rate of scarring was slightly overestimated because of the relatively short follow-up (7 mo), which might have been too limited to confirm that the demonstrated abnormalities were related to true scarring (22). However, combining imaging studies with routine clinical follow-up has been convenient. Finally, we chose a free period close to 6 mo before reinvestigation because some transient defects persist for up to 4 mo after APN before showing resolution on DMSA scans (5).

To avoid heterogeneity in sonographic findings, the same operator, using the same equipment, performed all studies in both the acute and the delayed phases of the study. In a recent study in which several operators performed sonography, interindividual variation occurred and the results of the study were less favorable to sonography (23). However, we emphasize that the same sonography equipment was used throughout our whole study to keep the DS results homogeneous.

In our study, DMSA 1 and DMSA 2 were performed using the same acquisition protocol, and reproducibility was ensured by double-blind analysis. SPECT was not performed. In a recent metaanalysis, SPECT DMSA resulted in a higher number of false-positive findings than did the established planar method (24). Moreover, SPECT is more expensive and, to yield clear images, requires that children be restrained at least twice as long (25 min vs. 12 min). As much as possible, we minimized the delay between sonography and DMSA to allow comparisons. However, DMSA 1 was performed 4 ± 3 d after DS 1. This difference in the timing of studies occurred because DMSA was not as available as was sonography, especially on weekends. However, we do not know whether some of the DMSA 1 studies with negative findings could, if performed sooner, have shown positive findings.

During the acute phase, the sensitivity of DS was 80% and the specificity was 81% in comparison with DMSA (Table 5). These findings confirm those of previous studies (8,16,17) and show the relative accuracy of sonography in detection of APN. The performance of sonography was slightly better for the right kidney (86% agreement on the right side vs. 75% on the left side), which is easier to image because of the hepatic window. For 10 patients satisfying the clinical and biologic criteria of APN, the results of DMSA 1 remained negative. DS 1 showed positive findings in 5 of these 10 patients (2 patients with isolated pyelitis of the left kidney, 1 with a focus of nephritis involving the left lower pole, and 2 with a focus of nephritis involving the left upper pole) and negative findings in 5 others. Such patients illustrate the statistical problem of lack of sensitivity in scintigraphy, which was thought to be the gold standard. Some of these patients might have had urinary tract infection without involvement of the renal parenchyma. In other cases, sonography might have revealed some lesions that remained invisible on DMSA. Rushton et al. (3) reported negative DMSA findings for piglets with APN. None had severe infection. The kidneys were grossly normal, and histopathologic findings consisted only of minimal microscopic foci of inflammatory cells.

Some kidneys with isolated pyelitis on sonography were shown to have a polar focus of decreased uptake on DMSA (4/7 kidneys), but pyelitis could not, of course, be visualized. We believe that sonography and DMSA do not strictly show the same disorder and that they can sometimes be complementary. Because DMSA scan sensitivity is not 100%, normal DMSA findings do not fully exclude upper urinary tract disease. No scarring developed in kidneys affected with isolated pyelitis.

Seven months later, 45 children were reexamined. Scarring had developed in 51%, and this rate can be compared with the 41% rate reported by Rushton and Majd (5). The results of imaging did not confirm our hypothesis: In fact, positive DS 1 results do not accurately predict the development of scarring. Moreover, in the acute phase, no other clinical, biologic, or imaging indicator could predict the patients in whom scarring would develop after infection. However, a definite conclusion can be drawn about children with normal DMSA findings during the acute phase, because in none of them did scarring subsequently develop. The difference between the 100% negative predictive value of DMSA 1 and the 75% negative predictive value of DS 1 stemmed from only 2 patients with false-negative DS 1 results (Table 6).

The development of scarring was not related to the presence of vesicoureteral reflux. This point agrees with previous reports. Rushton et al. (4) postulated that after APN had occurred, the development of scarring was independent of the presence or absence of reflux.

Children who have reflux after urinary tract infection usually receive antibiotic prophylaxis. The effect of such treatment in children with high-grade reflux (greater than grade III) is under evaluation. The treatment has been thought to decrease the prevalence of end-stage renal disease (25). However, given the results of this study, we postulate that management of children with no reflux or with reflux less than grade III could perhaps be changed. Because the presence of reflux does not correlate with the development of scarring, reflux should no longer remain the only reason to start antibiotic prophylaxis. In children with normal DMSA findings during APN, antibiotic prophylaxis does not seem useful, regardless of whether low-grade reflux is present. A further postulate is that patients with negative DMSA and DS findings would not need to undergo VCUG.

In contrast, children with abnormal DMSA findings during APN could be considered at risk of scarring. Adequate follow-up, including sonography, to assess renal growth and DMSA to detect scarring should be considered. Also, in these children, antibiotic prophylaxis may be useful even if reflux is absent.

CONCLUSION

A high rate of scarring after infection (51%) was found at an average follow-up interval of 7 mo. DS and DMSA were concordant in 81% of kidneys and 86% of patients for the diagnosis of APN. Therefore, DS can be considered relatively reliable in the acute phase, and the sensitivity should increase still further with new equipment. However, the predictive values of DS were relatively low, and it cannot yet be recommended as an accurate predictor of scarring. The positive predictive value of DMSA appeared insufficient as well. However, in no child with normal DMSA findings during APN did scarring develop. This study raises the question of whether the decision to use antibiotic prophylaxis in children with urinary tract infection should be based on the presence of scintigraphic abnormality rather than only on the presence of low-grade vesicoureteral reflux.