Pediatr Radiol (2002) 32: 849–852 DOI 10.1007/s00247-002-0784-6

Michael R. Ditchfield Dianne Summerville Keith Grimwood David J. Cook Harley R. Powell Robert Sloane Terrance M. Nolan John F. de Campo

Received: 3 September 2001 Accepted: 22 May 2002 Published online: 3 August 2002 Ó Springer-Verlag 2002 Presented at IPR, Paris, May 2001

M.R. Ditchfield (&) Æ D. Summerville D.J. Cook Æ J.F. de Campo Department of Radiology, Royal Children’s Hospital, Melbourne 3052, Australia E-mail: ditchfi[email protected] Tel.: +61-3-93455237 Fax: +61-3-93455284 K. Grimwood Æ R. Sloane Æ T.M. Nolan Department of General Paediatrics, Royal Children’s Hospital, Melbourne, Australia H.R. Powell Department of Nephrology, Royal Children’s Hospital, Melbourne, Australia K. Grimwood Æ T.M. Nolan Department of Paediatrics, University of Melbourne, Melbourne, Australia

ORIGINAL ARTICLE

Time course of transient cortical scintigraphic defects associated with acute pyelonephritis

Abstract Background: Acute pyelonephritis is distinguished from renal scarring using repeat cortical scintigraphy. The defects of acute pyelonephritis resolve, while those of scars persist. Objective: To determine the duration of reversible cortical defects following acute pyelonephritis and the time interval required to differentiate infection from scars. Materials and methods: An observational prospective study of 193 children (386 kidneys) aged less than 5 years following their first proven urinary tract infection (UTI). Renal cortical scintigraphic defects were detected in 112 (29%) kidneys within 15 days of diagnosis. Of these, 95 underwent repeat renal cortical scans 2 years after the UTI, including 50 with additional scans performed within 2–6 months of infection. Results: Of the 50 kidneys undergoing a second renal cortical scan within 2–6 months of the first UTI, 22 (44%) had persistent

Introduction A limitation of renal cortical scintigraphy is the potential difficulty in distinguishing acute pyelonephritis from renal scarring. Patel et al. [1] proposed criteria to distinguish between acute pyelonephritis and scarring from the appearance of the cortical defect, which is useful in the acute phase. However, the most reliable means of differentiating a scar from acute pyelonephritis is by serial renal cortical scintigraphy. Transient defects

defects. A third scan was performed on 17 (77%) kidneys after 2 years, by which time defects had resolved in another 8 (47%) kidneys. The predictive value of defects detected within 2–6 months of UTI representing scars is 53% (95% CI 28, 77). Overall, nine (18%) kidneys with initial renal cortical abnormalities had permanent defects. In the 45 kidneys undergoing a second cortical scan more than 6 months after the UTI, 11 (24%) had persistent defects. None of the 95 kidneys undergoing serial scans developed new or larger defects. Conclusions: Renal scars may not be reliably diagnosed by cortical scintigraphy performed within 6 months of UTI because the inflammatory lesions may not have fully resolved. Keywords Kidney Æ Pyelonephritis Æ Scar Æ Scintigraphy

indicate acute pyelonephritis, while persistent defects represent scars [2, 3]. Defects that persist can either be from scars present before the first renal cortical scintigram or from acute pyelonephritis that has evolved into a new scar. These can only be distinguished if an earlier study was performed before the episode of acute pyelonephritis. It is often recommended that there be a delay of at least 3 months following urinary tract infection (UTI) to allow resolution of parenchymal inflammation before

850

performing cortical scintigraphy to identify permanent defects [4]. The ideal time, however, for performing a follow-up study when a defect is discovered by renal cortical scintigraphy after a UTI has not been determined. This is important because if cortical scintigraphy is performed too early, a reversible defect from acute pyelonephritis may be falsely attributed to a permanent scar. In this study, we have attempted to describe the time course for reversible cortical defects associated with acute pyelonephritis by sequential scintigraphic examinations over 2 years.

Materials and methods Original cohort As previously described [5], 193 consecutive children (386 kidneys) aged younger than 5 years presenting to the emergency department with their first culture-proven UTI were recruited prospectively into an observational cohort study. Infection was defined by growth of at least 105 colony-forming units per millilitre of a single bacterial species from midstream or catheter specimens, or any pure growth of bacteria from a suprapubic aspirate of urine [6]. Treatment of the UTI was at the discretion of the attending physician. Children older than 6 months of age who were not vomiting and not acutely unwell received oral antibiotics. Infants younger than 6 months of age or those with a toxic appearance were hospitalised to receive parenteral amoxicillin and gentamicin until afebrile for at least 24 h. They were then discharged to complete a 7-day course of oral antibiotics, as determined by antibiotic susceptibility results. Those managed as out-patients received 7 days of trimethoprim-sulfamethoxazole (TMP-SMX). Following treatment, all received a single evening dose of either TMP-SMX or nitrofurantoin as prophylaxis until seen within 4 weeks of diagnosis by one of three study paediatricians. All children underwent imaging of the urinary tract, including renal scintigraphy, within 15 days of diagnosis. Children with a previous UTI or urinary tract obstruction were excluded. Approval was granted by the Royal Children’s Hospital Ethics in Human Research Committee. Study kidneys Acute renal cortical defects were detected in 112 (29%) of 386 kidneys from 86 (43%) of 193 children (81 boys, 112 girls; mean age 15 months, range 0.8–59 months); 65 children (34%) were hospitalised. Vesicoureteric reflux (VUR) was present in 41 (37%) of 112 kidneys with cortical defects. Of the 86 children with 112 renal defects, a randomly chosen subgroup of 38 possessing 50 kidneys with defects underwent additional scans within 6 months of diagnosis. All those with initial renal defects underwent repeat renal scintigraphy approximately 24 months after diagnosis of their UTI. Medical follow-up The children with VUR continued prophylactic antibiotic therapy for at least 3 months, until they were fully continent of urine and able to provide a weekly early morning urine sample for testing. The children with renal cortical defects, but without VUR, received continuous antibiotic therapy for a minimum of 3 months. They were followed up at 3-month intervals in the specialist clinic for at least 2 years.

Imaging protocols The children underwent renal sonography [5], a standard micturating cystourethrogram [7] and renal cortical scintigraphy that included high-resolution tomography [8]. A Starcam 400ACT gamma camera (General Electric Medical Systems, Milwaukee, Wis., USA) with high-resolution collimation was used and dimercaptosuccinic acid was administered, followed by delayed imaging 3 h later. A single, posterior-view delayed image of the kidneys was acquired for 300 K counts, followed by a 180° posterior tomographic acquisition of 32 steps, each of 30 s. Tomographic data were reconstituted by a Butterworth filter (0.7 or 0.8) and displayed as transverse, coronal, and sagittal tomograms. A nuclear medicine physician without knowledge of other imaging results graded the findings on the renal cortical scan from 0 to 3 (no defect, possible, probable and definite defects). For analysis, defects graded 0 or 1 were considered normal and those graded 2 or 3 were considered abnormal. When more than one defect was present, the kidney was graded by the most severe defect. When renal scintigrams were repeated within 6 months of the UTI, the children were taking antibiotic prophylaxis and were free of infection. To ensure that transient defects associated with subsequent UTI were not detected, the minimum interval between any subsequent UTI and the final follow-up renal cortical scintigraphy was 6 months. The time course of resolution of acute pyelonephritis on cortical scintigraphy was determined by examining those children with initial cortical defects who had renal cortical scintigraphy repeated on more than two occasions. The positive predictive value and specificity for renal scarring with corresponding 95% confidence intervals (CI) was determined for renal cortical scintigraphy performed at various delays after the diagnosis of the initial UTI.

Results Of the 112 (29%) kidneys with a cortical defect detected at the initial renal cortical scintigram, 95 (85%) had this investigation repeated on at least one occasion, including 50 kidneys where the scan was repeated within 2–6 months of the UTI. In these 50 kidneys with initial defects, 36 (72%) of the defects resolved and therefore represent acute pyelonephritis, 9 (18%) persisted and 5 (10%) did not have a scan at 2 years, meaning that the nature of these defects remains unknown. Even if the five kidneys lost to follow-up had persistent defects, at most, 14 (28%) kidneys with initial renal cortical abnormalities had permanent defects. Figure 1 shows the resolution of renal cortical defects over the 2-year period. Of the kidneys that underwent a second follow-up cortical scan more than 6 months after the UTI, 11 (24%) of 45 had a persistent defect. None of the 95 kidneys that underwent at least a second scan developed a new or larger defect. Table 1 demonstrates in more detail the results of the renal cortical scans performed between 2 and 6 months. A scan at 3 months with a defect has a positive predictive value for true renal scarring of 64% (95% CI 31, 89) and a specificity of 80% (95% CI 56, 94). If the timing of the final scan was 4 months, the positive predictive value

851

for true renal scarring was 40% (95% CI 5, 85) with a specificity of 73% (95% CI 39, 94). The small numbers in this study compromise the analyses and lead to broad confidence intervals. Combining the data for the 2–6month period following UTI, the positive predictive value that a defect detected in this period represents a scar is 53% (95% CI 28, 77) and the specificity is 78% (95% CI 61, 90).

Fig. 1 Resolution of defects at renal cortical scans performed between 2 and 6 months after urinary tract infection and then repeated 24 months later Table 1 Resolution of defects at renal cortical scans performed between 2 and 6 months after urinary tract infection (UTI) and then repeated 24 months later

Discussion This study has demonstrated that reversible defects may be present as long as 5 months after the initial infection. It is possible that reversible defects may persist for a longer period; however, this is difficult to establish without performing multiple cortical scans in a large group of children. This would not be ethical because of the invasive nature and radiation dose of each study. In our study, 47% of defects present at follow-up between 2 and 6 months resolved and therefore represented acute pyelonephritis rather than a scar. In other words, the positive predictive value that a defect detected in this period represents a scar is only 53%. A comparable study in Scandinavian children [9] reported cortical defects in 85% of children at the time of the initial radionuclide scan, by 5 months 58% had defects and after 2 years only 36% had defects present. It found that 44% of cortical scintigraphic defects detected between 3 and 5 months after UTI resolved by 2 years, a finding similar to our own. All children were in-patients, and presumably more likely to have acute pyelonephritis than our subjects recruited from an ambulatory setting, which would explain the higher incidence of initial cortical defects. These findings contrast with earlier reports of persistent defects being present in 50–80% of children following acute pyelonephritis [2, 10, 11, 12]. The follow-up investigations for many of these patients occurred within 6 months of their UTI. It is possible that many defects were still transient following the episode of acute pyelonephritis and may have resolved had there been a longer delay in renal imaging. The prevalence of renal scars following acute pyelonephritis may have been over-estimated in the medical literature by researchers performing renal cortical scintigraphy too soon after the original UTI and at a time when all the inflammatory lesions had not resolved. In the absence of adequate data there is conflicting advice in the literature over the timing of follow-up renal cortical scintigraphy to detect permanent defects and scars in children with a recent UTI. Some experts have suggested as little as 2 [13] or 3 months [4, 14], or as long as 6 [11, 15] or 12 months [16]. Although small numbers and corresponding broad confidence intervals limit the

Second cortical scan

Third cortical scan

Timing of 2nd scan after UTI

No. with 2nd scan

Defect on 2nd scan

No. with 3rd scan

2 months 3 months 4 months 5 months 6 months Total

2 29 16 2 1 50

0 13 8 1 0 22

0 11/13 (85%) 5/8 (63%) 1 (100%) 0 17

(45%) (50%) (50%) (44%)

Defect on 3rd scan

Defect resolved on 3rd scan

7/11 (64%) 2/5 (40%) 0/1 (0%)

4/11 (36%) 3/5 (60%) 1/1 (100%)

9 (53%)

8 (47%)

852

results of our study, the findings are remarkably similar to reports from Scandinavia [9]. Moreover, the proportion of children with persistent defects detected by a second scan performed after 6 months is comparable to those who had scans performed both within 6 months and 2 years of their infection. These data suggest that renal cortical scintigraphic defects following acute

pyelonephritis may be reversible for a longer period than previously believed. A defect detected less than 6 months after a UTI cannot be assumed to be permanent. Acknowledgements This study was supported by a grant from the Murdoch Children’s Research Institute.

References 1. Patel K, Charron M, Hoberman A, et al (1993) Intra- and interobserver variability in interpretation of DMSA scans using a set of standardized criteria. Pediatr Radiol 23:506–509 2. Benador A, Benador N, Slosman DO, et al (1994) Cortical scintigraphy in the evaluation of renal parenchymal changes in children with pyelonephritis. J Pediatr 124:17–20 3. MacKenzie JR, Fowler K, Hollman AS, et al (1994) The value of ultrasound in the child with an acute urinary tract infection. Br J Urol 74:240–244 4. Goldraich NP, Goldraich IH (1995) Update on dimercaptosuccinic acid renal scanning in children with urinary tract infection. Pediatr Nephrol 9:221– 226 5. Ditchfield MR, de Campo JF, Cook DJ, et al (1994) Risk factors in the development of early renal cortical defects in children with UTI. AJR 162:1393– 1397

6. Eichenfeld HF (1986) Some aspects of the diagnosis and management of urinary tract infection in children and adolescents. Pediatr Infect Dis J 5:760–765 7. Ditchfield MR, de Campo JF (1993) How to do it: the MCU in children. Aust Radiol 37:69–72 8. Traisman ES, Conway JJ, Traisman HS, et al (1986) The localisation of urinary tract infection with 99mTc glucoheptanate scintigraphy. Pediatr Radiol 16:403–406 9. Jakobsson B, Svensson L (1997) Transient pyelonephritic changes on 99mTechnetium-dimercaptosuccinic acid scan for at least five months after infection. Acta Paediatr 86:803–307 10. Jakobsson B, Nolstedt L, Svensson L, et al (1992) 99mTechnetium-dimercaptosuccinic acid scan in the diagnosis of acute pyelonephritis in children: relation to clinical and radiological findings. Pediatr Nephrol 6: 328–334 11. Lavocat MP, Granjon D, Guimpied Y, et al (1998) The importance of 99TcmDMSA renal scintigraphy in the followup of acute pyelonephritis in children: comparison with urographic data. Nucl Med Comm 19:703–710

12. Verber IG, Meller ST (1989) Serial 99mTc dimercaptosuccinic acid (DMSA) scans after urinary infections presenting before the age of 5 years. Arch Dis Child 64:1533–1537 13. Coulthard MG, Lambert HJ, Keir MJ (1997) Occurrence of renal scars in children after their first referral for urinary tract infection. BMJ 315:918– 919 14. Barry BP, Hall N, Cornford E, et al (1998) Improved ultrasound detection of renal scarring in children following urinary tract infection. Clin Radiol 53:747–751 15. Hoberman A, Wald ER, Hickey RW, et al (1999) Oral versus initial intravenous therapy for urinary tract infections in young febrile children. Pediatrics 104:79–86 16. Smellie JM, Rigden SP (1995) Pitfalls in the investigation of children with urinary tract infection. Arch Dis Child 72:251–258