Renal cortical scintigraphy in the diagnosis of acute pyelonephritis

Renal Cortical Scintigraphy in the Diagnosis of Acute Pyelonephritis Massoud Majd and H. Gil Rushton Comparative clinical studies have shown renal cortical scintigraphy, using technetium-ggm (~'Tc)-Iabeled glucoheptonate or dimercaptosuccinic acid (DMSA), to be significantly more sensitive than either intravenous pyelography or renal sonography in the diagnosis of acute pyelonephritis. However, due to uncertainties about the diagnostic accuracy of the clinical and laboratory parameters used in these studies, true sensitivity of renal cortical scintigraphy was unknown. Therefore, w e evaluated the accuracy of [ ~ F c l D M S A scintigraphy in the diagnosis of experimentally induced acute pyelonephritis in piglets using strict histopathologic criteria as the standard of reference. The sensitivity and specificity of the DMSA scan for the diagnosis of acute pyelonephritis were 91% and 99%, respectively, with an overall 97% agreement between the scintigraphic and histopathologic findings. Based on the results of this experimental study, we used the ["mTc]DMSA scan as the standard of reference for the diagnosis of acute pyelonephritis, and conducted a prospective clinical study of 94 children hospitalized w i t h the diagnosis of acute febrile urinary tract infection (UTI). The aims of this study were (1) to determine the relationship among vesicoureteral reflux, P-fimbriated Escherichia co/i, acute pyelonephritis, and renal scarring, and (2) to evaluate the diagnostic reliability of the clinical and laboratory parameters commonly used in the diagnosis of acute pyelonephritis. We documented acute pyelonephritis in 62 (66%) of 94 patients. Vesicoureteral reflux was demonstrated in 29 (31%) of the total group and in only 23 (37%) of 62 patients with acute pyelonephritis. The prevalence of P-fimbriae in the E c o i l isolates was 64% in the patients with acute pyelonephritis and 78% in those

w i t h a normal DMSA scan. Even in patients w i t h o u t reflux, P-fimbriae were found in 71% of isolates from the patients w i t h acute pyelonephritis and in 75% of those w i t h a normal renal scan. Follow-up DMSA scans were obtained in 33 patients with acute pyelonephritis in 38 kidneys. We found complete resolution of the acute inflammatory changes in 58% of the involved kidneys and renal scarring in the remaining 42%, including 40% of the kidneys associated w i t h reflux and 43% of those w i t h o u t reflux. The results of these experimental and clinical studies show the following: (1) [m"Tc]DMSA renal cortical scintigraphy is a highly sensitive and reliable technique for the diagnosis of acute pyelonephritis; (2) the diagnosis of acute pyelonephritis in children based on clinical and laboratory observations is unreliable; (3) acute pyelonephritis in the absence of reflux is common; (4) the presence of P-fimbriae alone does not fully explain the pathogenesis of acute pyelonephritis in the absence of reflux; (5) although high grades of reflux may be a risk factor for acute pyelonephritis, the risk of pyelonephritis in patients w i t h lower grades of reflux is the same as in those with no reflux; and (6) once acute pyelonephritis occurs, subsequent renal scarring is independent of the presence or absence of reflux. We conclude that DMSA renal cortical scintigraphy is a valuable diagnostic tool for investigating UTI, particularly in children. More precise diagnosis of acute pyelonephritis allows for newer insights into the pathophysiology of the disease and prevention of its sequelae, and in the future it may change the approach to the imaging evaluation and management of children with UTI. Copyright 9 1992 by W.B. Saunders Company

'rinary tract infection (UTI) may be limited U to the bladder (cystitis) or may involve the upper collecting systems (ureteritis or pyelits)

failure. Hypertension during adolescence or early adulthood has been reported in 10% to 18% of patients with renal scarring. 1"2The risk appears to be greater in patients with multifocal versus unifocal scarring. The remaining undamaged nephrons in the more severely scarred kidneys are subjected to hyperfiltration, which may result in focal segmental glomerulosclerosis and the development of proteinuria. 35 Once proteinuria occurs, there is usually an inexorable decline in renal function ultimately leading to renal failure. 6"7Renal scarring associated with vesicoureteral reflux accounts for as many as 10% to 20% of all patients with end-stage renal disease. ~~ Clinical and experimental studies have demonstrated that renal scarring can be prevented or diminished by early diagnosis and rigorous treatment of acute pyelonephritis. "13

or the renal parenchyma (pyelonephritis). Acute pyelonephritis is a major cause of morbidity in children with UTI, and it can result in irreversible renal scarring. Well-recognized sequelae of pyelonephritic scarring include hypertension, hyposthenuria, proteinuria, and chronic renal From the Departments of Radiology, Urology, and Pediatrics, Children's National Medical Center and The George Washington UniversitySchool of Medicine, Washington, DC. Address reprintrequests to Massoud Maid, MD, Department of Diagnostic Imaging and Radiology, Children's National Medical Center, 111 Michigan Ave, N~, Washington, DC 20010. Copyright ~ 1992 by W.B. Saunders Company 0001-2998/92/2202-0004/$05.00/0 98

Seminars in Nuclear Medicine, Vol XXII, No 2 (April), 1992: pp 98-111

ACUTE PYELONEPHRITIS

Therefore, accurate diagnosis of acute pyelonephritis has significant clinical relevance. PATHOGENESIS OF ACUTE PYELONEPHRITIS AND RENAL SCARRING

Experimental studies in primates suggest that the same acute inflammatory response that is responsible for the eradication of bacteria is also responsible for the damage to renal tissues and subsequent renal scarring, an The initiating event is bacterial innoculation of the renal parenchyma, which elicits both immune and inflammatory responses. Whereas the immune response can be elicited by either live or heatkilled bacteria, the acute inflammatory response occurs only following inoculation by live bacteria. Because heat-killed bacteria do not cause renal scarring, it appears that it is the acute inflammatory response that is more important in the subsequent development of permanent renal damage. 15 The inflammatory response of pyelonephritis is triggered by complement activation from bacterial lipopolysaccharides. This leads to chemotactic migration of granulocytes to the sites of infection. The offending bacteria are phagocytized by the granulocytes, initiating a sequence of events. Direct bacterial killing by the granulocytes occurs. Simultaneously, toxic enzymes, such as lyzozymes, are released both within the granulocytes and into the lumen of the tubules. At the same time, the respiratory burst occurs, a phenomenon universal to acute inflammatory responses. This releases superoxide that generates oxygen radicals, which are toxic not only to the bacteria, but also to the granulocytes and surrounding tubular cellsJ 6 This results in tubular cell death and extension of the inflammatory process into the interstitium. At the same time, intravascular granulocyte aggregation occurs, which, along with edema, creates focal ischemia. 17It is the combination of interstitial damage from both toxic enzymes and ischemia that ultimately leads to renal scarring. The role of vesicoureteral reflux in the pathogenesis of renal scarring is controversial. The common assumption that vesicoureteral reflux is a prerequisite for renal scarring is underscored by the popular term reflux nephropathy, which has been used interchangeably with

99

chronic atrophic pyelonephritis and renal scarring. This assumption has been based on the older studies of patients with known vesicoureteral refluxJ s'a9However, experimental studies in piglets have shown convincingly that, in the presence of vesicoureteral reflux and normal voiding pressures, renal scarring occurs only when urinary infection is present. 2~Reflux in the absence of infection causes renal scarring only when bladder outlet resistance is increased so that obstruction, not reflux, is the pathophysiological explanation for renal damage. The critical role that infection plays in the pathogenesis of renal scarring associated with vesicoureteral reflux is also evident from the results of long-term nonoperative management of children with known reflux. These studies have shown that subsequent progressive renal scarring can be successfully prevented by keeping the patient free of infection. 2x,z2 DIAGNOSIS OF ACUTE PYELONEPHRITIS

The diagnosis of acute pyelonephritis traditionally has been made on the basis of clinical signs and symptoms of fever and flank pain or tenderness associated with pyuria and positive urine culture. However, accurate diagnosis using these parameters often is difficult, especially in children. 23Neonates and infants in particular frequently present with nonspecific clinical findings. Various laboratory techniques--including erythrocyte sedimentation rate (ESR)r serum C-reactive protein levels, urinary beta-2 microglobulin excretion, antibody-coated bacteria in the urine sediment, and urinary lactic dehydrogenase assays--have been reported. For a variety of reasons, including unreliability and/or nonavailability, these have not gained widespread acceptance for the evaluation of children with acute urinary tract infectionfl Splitureteral catheterization or bladder washout studies may accurately localize infection to the upper or lower urinary tract, but they are invasive and do not differentiate pyelitis from parenchymal involvement. Furthermore, none of these techniques provides any information regarding renal function or extent of parenchymal involvement by acute infection. Several imaging techniques have been evaluated as a means of differentiation of acute pyelonephritis from lower urinary tract infec-

100

tion. Intravenous urography (IVU) and renal sonography both have a very low sensitivity for the diagnosis of acute pyelonephritis. Computed tomography (CT) is probably a sensitive and effective technique for documenting the nature and extent of parenchymal involvement and particularly for the evaluation of the perinephric space. However, its routine use in the initial evaluation and follow-up examinations of children with UTI is not practical. '-5'26The role of magnetic resonance imaging in the diagnosis of acute pyelonephritis is yet to be defined. 27 Radionuclide imaging procedures using gallium-67 citrate or indium-Ill-labeled leukocytes may be reliable in the diagnosis of acute pyelonephritis. 28"z9However, these imaging procedures result in a high radiation absorbed dose, require 24 to 48 hours to perform, and, more importantly, do not provide any information about the function and morphology of the kidneys. Therefore, they are not suited for routine use in the evaluation of children with febrile UTI. Recent clinical studies have shown renal cortical scintigraphy using technetium-99m (99mTc)-labeled glucoheptonate or dimercaptosuccinic acid (DMSA) to be significantly more sensitive than IVU and renal sonography in the detection of renal parenchymal involvement. 3~ In these clinical reports the various imaging modalities have been retrospectively compared with clinical signs and symptoms and routine laboratory studies. However, these same clinical and laboratory findings have been shown to have poor correlation to direct localization studies, such as the bladder washout test. 23 Therefore, true sensitivity and specificity of renal cortical scintigraphy was unknown. This prompted us to evaluate the accuracy of [99mTc]DMSA renal cortical scintigraphy in the diagnosis of acute pyelonephritis in an animal model using strict histopathologic criteria as the standard of reference.

MAJD AND RUSHTON

similar to humans, and it is a well-established animal model for the study of pyelonephritis. The experimental design and the surgical procedures were described in detail in a previous report. 33 Briefly, unilateral (right or left) vesicoureteral reflux was created randomly by incising the roof of the intramural ureter. After a 2-week healing period a cystogram was obtained to confirm the presence of vesicoureteral reflux (Fig 1). Urinary infection subsequently was induced by the introduction into the bladder of a broth culture of Escherichia coli. Technetium99m DMSA scans were obtained at 1 week or 2 weeks after the introduction of bacteria into the bladder (Fig 2). The images were obtained in posterior and posterior-oblique projections using a small-field-of-view camera equipped with a converging collimator. Focal areas of decreased cortical uptake of DMSA without loss of volume in kidneys with or without enlargement or diffuse decreased uptake in enlarged kidneys were interpreted as acute pyelonephritis. The piglets were then killed, and the kidneys

[99mTc]DMSA Scintigraphy in the Diagnosis of Experimentally Induced Acute Pyelonephritis We evaluated the sensitivity and specificity of [99mTc]DMSA renal cortical scintigraphy in the detection and localization of experimentally induced acute pyelonephritis in piglets. This animal has multipapillary renal architecture

Fig 1. Cystogram of a piglet demonstrates unilateral vesicoureteral reflux.

ACUTE PYELONEPHRITIS

Fig 2. Experimental acute pyelonephritis in a piglet. Technetium-99m DMSA scan obtained 1 week after the introduction of E coil into the bladder shows a large area of decreased uptake of DMSA in the upper pole of the right kidney. There is also a small area of subtle decreased uptake in the lateral aspect of the lower pole of the right kidney. The left kidney is normal.

were examined for histopathologic evidence of acute pyelonephritis characterized by the presence of intratubular neutrophils and interstitial infiltrates of mononuclear cells and neutrophils (Fig 3). The scintigraphic findings were then compared with the histopathological findings in a blinded fashion.

101

Based on our initial study and subsequent additional animals, 29 kidneys were subjected to vesicoureteral reflux of infected urine. Among these, 22 had histopathologic evidence of acute pyelonephritis. DMSA scans showed scintigraphic evidence of acute pyelonephritis in 20 of 22 kidneys (sensitivity = 91%). The 2 kidneys in which inflammation was not detected on the scans had only minimal microscopic foci of inflammatory cells and were grossly normal. There were no-false positive scans in any of the 29 experimental kidneys (specificity = 100%). DMSA scans also correctly correlated with pathological findings in all but 2 of 29 contralateral kidneys not subjected to reflux. There was a minimal microscopic focus of inflammatory ceils in one kidney, which was not detected on the scan (false negative). In another kidney a focal defect on the scan, which proved to be due to a large caliceal diverticulum, was interpreted as acute pyelonephritis (false positive). To eval~ uate the reliability of the DMSA scan in the detection and localization of the individual pyelonephritic lesions, each kidney was arbitrarily divided into upper, middle, and lower zones. The scans correctly diagnosed the presence or absence of the lesions in 169 (97%) of 174 zones (sensitivity = 91%; specificity = 99%).

Fig 3. Experimental acute pyelonephritis in a piglet, (A) Gross photograph of the right kidney shows a large area of pallor involving the upper pole. There is also an area of pallor associated with bulging border in the lower lateral aspect of the kidney. (B) Light micrograph shows histological findings of acute pyelonephritis characterized by the presence of intratubular neutrophils and interstitial infiltrates of mononuclear cells and neutrophils.

102

Subsequent to our report, Parkhouse et al, using a similar experimental model, reported the results of their study in 33 piglets with a total of 37 refluxing ureters. Histopathologic evidence of acute pyelonephritis was noted in 27 of 37 kidneys involving a total of 64 zones: upper, middle, and lower. The scans correctly demonstrated the lesions in 52 zones of 24 kidneys. Therefore, in their study the sensitivity (true positive/true positive + false negative) of the DMSA scan for the detection of pyelonephritis in the kidneys subjected to reflux of infected urine was 89% (24 of 27) and for the detection of individual lesions was 81% (52 of 64). The specificity (true negative/true negative + false positive) was 100% for both the involved kidneys and the zones. 34 The results of these studies show that DMSA renal cortical scintigraphy is a highly sensitive and reliable technique for the detection and localization of experimental acute pyelonephritis. There is no reason to believe that the same degree of accuracy cannot be obtained in the diagnosis of acute pyelonephritis in humans. Therefore, at present the DMSA scan, although not perfect, appears to be the best clinically applicable standard of reference for the diagnosis of acute pyelonephritis and for the evaluation of sensitivity and specificity of other imaging techniques in clinical investigations.

MAJD AND RUSHTON

Technique of [9~Tc]DMSA Scintigraphy In clinical practice, we routinely image the kidneys 90 to 120 minutes after intravenous injection of [99mTc]DMSAin a dose of 50 ~Ci/kg body weight (minimum 300 ~Ci). A posterior image of both kidneys together, using a highresolution parallel-hole collimator, is obtained for the evaluation of the relative size and cortical uptake of the kidneys. Magnified posterior and posterior-oblique images of each kidney are also obtained using a pinhole collimator (Fig 4). Additional lateral views or single photon emission computed tomography images are obtained in selected cases.

Scintigraphic Patterns Normal scan. The cortical uptake of DMSA may appear homogeneous throughout the kidneys except for a lower concentration in the areas overlying the collecting systems. However, the high-resolution magnified images obtained with the state-of-the-art cameras, particularly with the use of pinhole collimator, often show heterogeneous uptake due to prominent cortical columns and better differentiation of the cortex from the medulla (Fig 4B). Flattening of the superolateral aspect of the upper pole of the left kidney due to splenic impression may be seen (Fig 5). Irregularities in the contour of the

A

ip Fig 4. Normal [m"Tc]DMSA renal cortical scan of a 9-month-old infant. (A) Posterior image obtained with a parallel-hole collimator and (B) magnified images of the kidneys in posterior and posterior-oblique projections using pinhole collimator show normal cortical uptake bilaterally. Note normal cortical columns.

ACUTE PYELONEPHRITIS

kidneys due to fetal lobulation may be present, but the cortical thickness and uptake in the areas under the indentations are normal (Fig 6). Recognition of these normal variants is essential in the interpretation of the scans. Acutepyelonephritis. Acute pyelonephritis is usually manifested as a single or as multiple areas of varying degrees of diminished cortical uptake of DMSA with or without bulging contour. The important feature of the acute lesion is that the defect in uptake is not associated with loss of volume. Although the majority of the lesions occur in the upper or lower poles, the midzone of the kidney is frequently involved (Figs 7 and 8). A less common scintigraphic pattern is one of diffuse decreased uptake in an enlarged kidney (Fig 9). Acute pyelonephritis may resolve completely, and the scan may return to normal within 6 months (Fig 10), or it may evolve into permanent damage and scar formation (Fig 11). Cortical scar A mature cortical scar is usually associated with contraction and loss of volume of the involved cortex. This may manifest as cortical thinning, flattening of the renal contour, or a wedge-shaped defect (Figs 11 and 12). The defect of a mature scar may gradually become more prominent due to growth of the surrounding normal cortex. The scintigraphic pattern of a maturing scar varies depending on the severity, age, and location of the lesion as well as the rate of the growth of the surrounding normal tissues.

103

Mechanism of Decreased Uptake of DMSA in Acute Pyelonephritis The pathophysiological mechanisms that account for the decreased uptake of DMSA in acute pyelonephritis are probably multifactorial. Cortical uptake of DMSA depends on renal blood flow and proximal tubular cell membrane transport function. Any pathological process that alters either or both of these parameters may result in focal or diffuse decreased uptake of DMSA. During acute inflammation, intratubular neutrophils release toxic enzymes and produce superoxide, causing direct damage not only to bacteria but to renal tubular cells a s w e l l . ~6 In an experimental study in primates, it has been demonstrated that ischemia, as indirectly evidenced by elevated renal vein renin, occurs early in inflammatory response to acute pyelonephritis. This has been attributed to intravascular granulocyte aggregation leading to arteriolar or capillary occlusion. ~7 In another study, microvascular changes in experimental acute pyelonephritis in pigs were studied using combined stereomicroscopic and microradiographic techniques. Focal ischemia in areas involved by the acute inflammatory response was evidenced by compression of glomeruli, peritubular capillaries, and vasa rectae, presumably from interstitial edema? 5 Thus, diminished uptake of DMSA in areas of acute inflammation probably reflects both focal tubular cell dysfunction and ischemia.

Fig 5. Splenic impression. (A) Technetium-99m DMSA scan and (B) longitudinal sonogram of the left kidney show marked flattening of the upper pole of the left kidney due to splenic impression, Note normal cortical uptake of DMSA in this region.

104

Fig 6. Fetal Iobulations of the kidneys in a 6-month-old infant. Note normal cortical thickness and DMSA uptake under the indentations.

We evaluated focal renal blood flow changes in experimentally induced acute pyelonephritis in a group of piglets using the microsphere technique. After obtaining DMSA scans, chromium-51-microspheres were injected into the left ventricle. Subsequently the piglets were killed, and multiple tissue samples were obtained from the diseased and nondiseased areas of the experimental kidneys and the normal contralateral kidneys. Using standard counting techniques the percentages of decreased uptake per gram of tissue in the samples from each experimental kidney relative to the average uptake in the contralateral normal kidney were calculated for both [99mTc]DMSA and SlCrmicrospheres. We found a significant decrease in blood flow in almost all of the pyelonephritic lesions but in none of the samples from the uninvolved areas of the same kidneys. There was also no evidence of decreased blood flow in

Fig 7. Unifocal acute pyelonephritis in a 2-year-old child. Note decreased uptake of DMSA in the upper pole of the right kidney without any evidence of loss of volume.

MAJD AND RUSHTON

the experimental kidneys with reflux that did not have any evidence of acute pyelonephritis. In approximately 40% of the pyelonephritic lesions the decreased uptake of DMSA was proportional to the decrease in blood flow. In the remaining 60%, DMSA uptake was much more severely decreased than corresponding decrease in blood flow. The results of this study suggest that ischemia is an early event that precedes tubular cell dysfunction. Therefore, the DMSA scan probably becomes positive early in the course of the disease before any significant tissue damage has occured. 36

DMSA Versus Glucoheptonate Technetium-99m DMSA is an excellent renal cortical imaging agent. Approximately 60% of the administered dose is tightly bound to the proximal tubular cells, and only a small amount is excreted slowly in the urine. DMSA allows visualization of renal parenchyma without interference from retained tracer in the collecting systems. Technetium-99m glucoheptonate is cleared by both glomerular filtration and cortical fixation. About 20% of the administered dose is firmly bound to the tubular cells. Therefore, delayed imaging provides visualization of the renal cortex similar to DMSA images. Most of the administered dose is excreted in the urine, allowing moderately good visualization of the renal collecting systems on the early images. This additional information provided by glucoheptonate makes it the agent preferred by some. However, the status of the collecting systems is best evaluated on the renal sonogram, which usually is obtained in conjunction with the renal scan. There is often some heptobiliary excretion of glucoheptonate, particularly in infants (Fig 13). Tracer in the gallbladder or small bowel may be superimposed on the right kidney. Concerns regarding higher radiation absorbed dose with DMSA have been expressed. 37 It is true that radiation to the cortex of the kidneys per milicurie of administered dose is three times higher with DMSA compared with glucoheptonate, but this is due to cortical fixation of a larger fraction of DMSA. Therefore, the administered dose of DMSA can be one third that of glucoheptonate (50 t~Ci/kg v 150 I~Ci/Kg body weight). With this adjustment in the administered dose, total DMSA

ACUTE PYELONEPHRITIS

105

Fig 8. Multifocal acute pyelonephritis in a 4-year-old child. (A) The posterior image and (B) the magnified posterior and posterior-oblique images demonstrate three areas of decreased DMSA uptake in the right kidney, The lesion in the lower pole is best seen on the right posterior-oblique image.

retained in the kidney and resultant renal radiation doses are equal to those of glucoheptonate, but, owing to the accumulation of a lesser amount of DMSA in the bladder, the gonadal radiation dose is significantly less. Therefore, we prefer to use DMSA, particularly in infants and young children.

The Role of Sonography in the Diagnosis of Acute Pyelonephritis Renal sonography is commonly used in the evaluation and management of UTI. It is a truly noninvasive imaging technique that is highly reliable for the detection of hydronephrosis and some of the congenital anomalies that may be

associated with UTI. It is also useful in the detection of renal abscesses, pyonephrosis, and abnormalities of the perinephric space. Changes secondary to acute pyelonephritis may also be recognized on the renal sonogram as either hyperechoic or hypoechoic foci, loss of corticomedullary differentiation, focal or diffuse renal enlargement, and mild to moderate dilatation of the renal pelvis. However, renal sonography has proved to have a low sensitivity for the detection of acute inflammatory changes of the renal cortex. In a prospective study, sonographic changes consistent with acute pyelonephritis were found in only 39% (22 of 57) of children with scintigraphically documented acute

A

Fig 9. Diffuse acute pyelonephritis in a 5-week-old infant. Note generalized decreased DMSA uptake in enlarged (swollen) left kidney.

Fig 10. Resolution of acute pyelonephritis in the upper pole of the left kidney of a 3-month-old infant. The interval between (A) the initial scan and (B) the follow-up scan w a s 16 months.

106

MAJD AND RUSHTON

C

B

D

l Fig 12. Cortical scars in a 9-year-old child. Note wedgeshaped defects in the lateral aspect of the upper pole of the left kidney and in the lateral aspect of the midzone of the right kidney.

Fig 11. Progression of acute pyelonephritis to cortical s c a r in a 5-year-old child, The posterior (A) and the left posterioroblique (B) images of the left kidney from the initial study demonstrate decreased uptake in the upper pole with preserved renal outline and without any evidence of loss of volume characteristic of acute pyelonephritis. The corresponding images from the follow-up study 1 year later (C and D) show contraction and loss of volume of the upper pole.

with vesicoureteral reflux, and that vesicoureteral reflux is a prerequisite for acquired renal scarring. However, recent studies have suggested that acute pyelonephritis in the absence of reflux is more common than previously thought, and that acquired renal scarring following urinary tract infection also occurs in the absence of demonstrable vesicoureteral ref l u x . 39"12 Acute pyelonephritis in the absence of

pyelonephritis. Furthermore, sonograms underestimated the number and extent of pyelonephritic lesions in the majority of cases compared with the DMSA scan. 38 This is in close agreement with the 45% positive rate for sonogram in 18 adults with CT evidence of acute pyelonephritis. 25Therefore, renal sonography should not be used as the primary imaging technique for the diagnosis of acute pyelonephritis. However, it is useful in characterizing the defects seen on the DMSA cortical scan and in detecting obstructive uropathies that may be associated with urinary tract infection. TECHNETIUM-99M DMSA SCINTIGRAPHY IN CLINICAL INVESTIGATIONS

It is commonly assumed that most cases of acute pyelonephritis in children are associated

Fig 13. Hepatobiliary excretion of m"Tc[glucoheptonate] in a 1g-month-old infant. Right posterior-oblique image shows tracer in the liver and the gallbladder.

ACUTE PYELONEPHRITIS

reflux has been attributed to bacterial adherence to urothelial cells related to the presence of P-fimbriae in certain strains of E coli. 43 The drawback to these retrospective clinical studies is that the diagnosis of acute pyelonephritis is based on the clinical signs and symptoms and laboratory findings of leukocytosis and elevated ESR, which may not be accurate. Based on the results of our experimental study, we used the [99~Tc]DMSA cortical scan as the standard of reference for the diagnosis of acute pyelonephritis and conducted a prospective clinical study. 44'45The aims of this study were (1) to determine the relationship among vesicoureteral reflux, P-fimbriated E coli, acute pyelonephritis, and cortical scarring, and (2) to evaluate the reliability of commonly used clinical and laboratory observations in the diagnosis of acute pyelonephritis. The study population consisted of 94 consecutive children admitted to Children's National Medical Center, Washington, DC with the diagnosis of an acute febrile culture-documented UTI. The patients ranged in age from 2 weeks to 18.9 years (mean, 3.0; median, 1.0 years). Fifty-eight were female. There were 54 black and 40 nonblack patients (35 patients were white). Renal sonography and [99mTc]DMSA renal cortical scintigraphy were carried out within 72 hours of admission. Cystography was performed in all patients before discharge. Clinical and laboratory parameters, including maximum daily temperature, presence of chills, lethargy, nausea, vomiting, anorexia, abdominal pain, costovertebral angle (CVA) tenderness, complete blood count, ESR, urine culture and blood culture, were recorded. In addition, urinary E coli isolates were analyzed for the presence of P-fimbriae. Follow-up DMSA scans were obtained in a subgroup of 33 patients with scintigraphically documented acute pyelonephritis. The DMSA scans showed acute pyelonephritis in 66% (62 of 94) of patients. Diffuse changes were present in 3, unifocal lesions in 32, and multifocal involvement in 27. Bilateral changes were noted in 12 patients. Acute pyelonephritis involved the upper pole of 55, lower pole of 39, and midzone of 24 kidneys. Cortical scarring was present in 14 patients (15%), 10 of whom also had acute pyelonephritis.

107

Vesicoureteral reflux was found in 31% (29 of 94) of the total group and in only 37% (23 of 62) of patients with evidence of acute pyelonephritis on DMSA scan. Reflux was unilateral in 15 and bilateral in 14 patients. Among 43 refluxing units, the grade of reflux was mild in 22, moderate in 15, and severe in 6. The prevalence of reflux was 15% (8 of 54) in blacks and 53% (21 of 40) in nonblacks. The prevalence of reflux was 37% (23 of 62) in the group of patients with scan evidence of acute pyelonephritis and 19% (6 of 32) in the group with normal scan findings [P = .112]. DMSA scan findings of acute pyelonephritis were present in 79% (23 of 29) of patients with reflux and in 60% (39 of 65) of those without demonstrable reflux (Table 1). Acute pyelonephritis was seen in 66% (14 of 21) of kidneys with either moderate or severe reflux, 32% (7 of 22) of those with mild (grade I to II) reflux, and 35% (51 of 145) of those with no reflux. The organisms isolated from the urine culture were E coli in 83, klebsiella species in 4, and pseudomonas species in 2 patients. Isolated examples of proteus, citrobacter, enterococcus, salmonella, and staphylococcus species were found in the remaining 5 patients. P-fimbriated E coli were found in 69% (48 of 70) of patients whose E coli isolates were studied. The prevalence of P-fimbriated E coli was 64% (30 of 47) in the group with acute pyelonephritis and 78% (18 of 23) in the group with normal scans [P = .343]. Among 54 patients without reflux, the prevalence of P-fimbriated E coli was 71% (24 of 34) in children with acute pyelonephritis and 75% (15 of 20) in those with a normal scan. However, among 47 patients with acute pyelonephritis, the prevalence of P-fimbriated E coli was 46% (6 of 13) in children with reflux and 71% (24 of 34) in those without reflux ]P = .222] (Table 2). Only 9% (7 of 76) of blood cultures were positive, all yielding E coli. Four of these occurred in patients who had acute pyelonephritis Table 1. Relationship of [SgmTc]DMSA Scan Findings and Vesicoureteral Reflux in 94 Patients DMSA Scan Reflux

+

-

Total

+

23

6

29

--

39

26

65

To~I

62

32

94

108

MAJD AND RUSHTON

Table 2. Relationship of P-fimbriae, Reflux, and D M S A Scan Findings in 70 Patients Scan (+) P-fimbriae Reflux(+) +

Total

6 7 13

Scan(+) Reflux(-)

Scan(-) Reflux(+)

Scan(-) Reflux(-)

Total

24

3

15

48

10 34

0 3

5 20

22 70

with no evidence of vesicoureteral reflux. Two patients had vesicoureteral reflux and acute pyelonephritis. The remaining patient had a normal cystogram and DMSA renal scan. The age of the patients with positive blood cultures ranged from 6 weeks to 10.4 years (mean, 2.4; median, 0.69).

Unreliability of Clinical and Laboratory Parameters in the Diagnosis of Acute Pyelonephritis The clinical diagnosis of acute pyelonephritis is controversial. The criteria used by different investigators vary greatly. The parameters commonly used in children to differentiate acute pyelonephritis from lower UTI include high fever, CVA tenderness, abdominal pain, chills, lethargy, vomiting, leukocytosis (white blood cell count > 12,000), elevated ESR ( > 25), lowered concentrating capacity, and presence of r e f l u x . 11'31'37 Our data show that a high percentage of children hospitalized with febrile UTI have evidence of acute pyelonephritis on [99mTc]DMSA renal scan. However, regression analysis of multiple clinical and laboratory parameters commonly used in the diagnosis of acute pyelonephritis, including the presence of reflux and P-fimbriae, did not demonstrate a substantial improvement in the accuracy for predicting parenchymal involvement, as evidenced by DMSA scan, beyond that achieved using fever alone. Therefore, renal parencymal involvement cannot be predicted on the basis of the clinical and laboratory parameters.

Relationship Between Vesicoureteral Reflux and Acute Pyelonephritis We found a higher prevalence of vesicoureteral reflux in patients with scan-documented pyelonephritis (37%) compared with patients with normal scans (19%), underscoring the role that vesicoureteral reflux may play in the pathogenesis of pyelonephritis in children.

Even so, vesicoureteral reflux was not found in the majority of our patients with febrile urinary infections, and the prevalence of reflux in these patients with scan-documented pyelonephritis (37%) did not vary significantly from the 35% prevalence of reflux generally reported by others for all children with symptomatic urinary tract infections. 4~This may be explained in part by the racial distribution in the study population, 57% of whom were black. The prevalence of reflux in black children with UTI has been reported to be significantly lower than in whites. 47 In this series, reflux occurred in 15% of blacks compared with 53% of nonblacks [P < .001]. Regardless of the influence of race on the prevalence of reflux, acute pyelonephritis, as documented by the DMSA scan in children with febrile UTI, is common even in the absence of demonstrable vesicoureteral reflux.

Relationship Among P-Fimbriated E Coli, Vesicoureteral Reflux, and Acute Pyelonephritis The pathogenesis of acute pyelonephritis in the absence of reflux is controversial. Some attribute this to hematogenous infection, particularly in neonates. 48 However, the low prevalence of positive blood cultures in the present study does not support this hypothesis. Others have attributed pyelonephritis in the absence of gross vesicoureteral reflux to ascending infection, emphasizing the importance of bacterial virulence factors such as adherence to urothelial cells. 43 Bacterial adherence of E coli to uroepithelial cells has been shown to be related to the presence of filamentous protein structures termed fimbriae or pili, specifically P-timbriae.49 These structures attach to specific globoseries glycolipid receptors located on the surface of uroepithelial cells? ~ In one study, P-fimbriated E coli were found in the urine of 94% of children with clinical acute pyelonephritis (defined by fever greater than 38~ ESR of at least 20, and/or C-reactive protein level of at least 10 mg/mL) compared with only 19% of children with cystitis and 14% with asymptomatic bacteriuria. 5z We identified P-fimbriated E coli isolates in 69% (48 of 70) of children hospitalized with febrile urinary tract infections. This is similar to the findings reported by others. 43 However, the prevalence of P-fimbriated E coli was less in patients with acute pyelonephritis

ACUTE PYELONEPHRITIS

(64%) than in those with normal DMSA renal scans (78%). Furthermore, even among our patients without reflux the prevalence of P-timbriated E coli was less in children with pyelonephritis (71%) than in those with a normal scan (75%). Experimental studies in primates have shown that, in the absence of reflux, inoculation of the bladder with P-flmbriated E coli resulted in pyelonephritis in 66% of animals. In contrast, pyelonephritis was not seen in any of the monkeys inoculated with non-P-fimbriated E coli. 53 In the present study, among 54 patients without reflux, acute pyelonephritis was documented in 62% (24 of 39) of patients infected with P-timbriated E coli and in 67% (10 of 15) of those infected with non-P-fimbriated E coil Among the patients with acute pyelonephritis, the prevalence of P-fimbriated E coli was higher in children without reflux (71%) than in those with reflux (46%; P = .222). Therefore, the presence of P-fimbriae alone does not fully explain the pathophysiology of renal parenchymal invasion by bacteria in the absence of reflux. Although the presence of P-fimbriae may be important in colonization of the upper collecting systems, data from the present study suggest that bacterial virulence factors other than P-fimbriae or host-defence mechanisms are more important in renal parenchymal invasion. Relationship Between Vesicoureteral Reflux and Postpyelonephritic Renal Scarring Many studies have reported the close association between postpyelonephritic renal scarring and vesicoureteral reflux. 46 We obtained follow-up DMSA scans in 33 of 62 patients with scintigraphically documented acute pyelonephritis in 38 kidneys, including 15 with and 23 without reflux. 45The interval between the initial and the follow-up scans ranged from 4 to 42 months (mean, 10.7). Follow-up scans showed complete resolution of the acute changes in 22 of 38 (58%) kidneys. New renal scars were seen in the remaining 16 (42%) kidneys. The new scars developed only at the sites corresponding exactly to the sites of acute changes on the initial scans. The most significant and unexpected finding in this study was that new scarring in kidneys without reflux (43%) was as common as in those with reflux (40%). These

109

findings provide convincing evidence that renal parenchymal infection, rather than vesicoureteral reflux, is the prerequisite for acquired renal scarring. Once bacteria have invaded the renal parenchyma, the inflammatory response appears similar, and the propensity for renal scarring in areas of acute inflammation is equally as great whether or not reflux is present. Currently, the indication for low-dose antibiotic prophylaxis in infants and children with UTI is usually based solely on the presence of vesicoureteral reflux regardless of its severity. Perhaps direct scintigraphic evidence of parenchymal involvement at the time of acute symptomatic UTI is a more important determinant of the need for prophylaxis. A child with a positive DMSA scan, regardless of the status of reflux, may be a more reasonable candidate than a child with low-grade reflux and a normal DMSA scan. Further prospective clinical studies are needed to determine the risk of recurrent pyelonephritis and its consequences in patients with pyelonephritis in the absence of reflux. SUMMARY

From the experimental and clinical data presented in this review, the following conclusions can be drawn: 1. Renal cortical scintigraphy is a highly sensitive and reliable technique for the diagnosis of renal parenchymal infection, and it should be considered as the standard of reference for the diagnosis of acute pyelonephritis. 2. Decreased uptake of DMSA in acute pyelonephritis is due to both ischemia and tubular cell dysfunction. The DMSA scan probably becomes positive in the early phase of the inflammatory response as a result of ischemia and before any significant tubular damage has occurred. 3. Commonly used clinical and laboratory parameters are not reliable for thr diagnosis of acute pyelonephritis in children and its differentiation from the upper tract infection without parenchymal involvement. 4. Acute pyelonephritis in the absence of reflux is common in children. 5. The presence of P-firnbriaeted E coli alone does not fully explain the pathogenesis of

110

MAJD AND RUSHTON

acute pyelonephritis in the absence of reflux. Although the presence of P-funbriae is important both for the colonization of the upper urinary tract and induction of a febrile response, some other bacterial virulence and/or host defense factors are more important in parenchymal invasion by the bacteria. 6. Although higher grades of reflux appear to be a risk factor for acute pyelonephritis, the risk of pyelonephritis in the kidneys

with mild reflux (grade I to II) is the same as in those without reflux. 7. Acquired renal scarring only occurs at sites corresponding to previous areas of acute pyelonephritis. 8. The acute parenchymal inflammatory changes are reversible and do not lead to renal scarring in the majority of cases. 9. Once acute pyelonephritis occurs, subsequent renal scarring is independent of the presence or absence of reflux.

REFERENCES 1. Holland NII, Kotehen T, Bhathena D: Hypertension in children with chronic pyelinephritis. Kidney Int 8:S-243251, 1975 (suppl) 2. Wallace DMA, Rothwell DL, Williams DI: The long term follow-up of surgically treated vesicoureteral reflux. Br J Uro150:479-484, 1978 3. Kincaid-Smith P: Glomerular lesions in atrophic pyeionephritis and reflux nephropathy. Kidney Int 8:S-81-83, 1975 (Suppl) 4. Bhathena DB, Weiss JH, Holland Nil, et al: Focal and segmental glomerular sclerosis in reflux nephropathy. Am J Med 68:886-892, 1980 5. Steinhardt GF: Reflux nephropathy. J Uroi 134:855859, 1985 6. Torres VE, Velosa JA, Holley KE, et al: The progression of vesicoureteral reflux nephropathy. Ann Int Med 92:776-784, 1980 7. Berger RE, Ansell JS, Shurtleff DB, et al: Vesicoureteral reflux in children with uremia: Prognostic indicators for treatment and survival. JAMA 246:56-59, 1981 8. Bailey RR: End-stage reflux nephropathy. Nephron 27:302-306, 1981 9. Senekjian HO, Suki WN: Vesicoureteral reflux and reflux nephropathy. Am J Nephrol 2:245-250, 1982 10. Jacobson SH, EklOf O, Eriksson CG, et al: Development of hypertension and uremia after pyeionephritis in childhood: 27 year follow up. Br Med J 299:703-706, 1989 11. Winberg J, Boligren I, I~llenius G, et ai: Clinical pyelonephritis and focal renal scarring: A selected review of pathogenesis, prevention, and prognosis. Pediatr Clin North Am 29:801-814, 1982 12. Ransley PG, Risdon RA: Reflux nephropathy: Effects of antimierobial therapy on the evolution of the early pyelonephritic scar. Kidney Int 20:733-742, 1981 13. Gtauser MP, Lyons JM, Braude ALl: Prevention of chronic experimental pyelonephritis by suppression of acute suppuration. J Clin Invest 61:403-407, 1978 14. Roberts JA: Pathogenesis of pyelonephritis. J Urol 129:1102-1106, 1983 15. Roberts JA, Dominique G J, Marlin LN, et al: Immunology of pyelonephritis in the primate model: Live versus heat killed bacteria. Kidney Int 19:297-305, 1981 16. Roberts JA, Ruth JK Jr, Dominique GJ, et al: Immunology of pyeionephritis in the primate model. V. Effect of superoxide dismutase. J Urol 128:1394-1400, 1982 17. Haack MB, Dowling KJ, Patterson GM, et al: Immu-

noiogy of pyelonephritis. VIII. E. coli causes granulocytic aggregation and renal ischemia. J Urol 136:1117-1122, 1986 18. Bailey RR: The relationship of vesico-ureteric reflux to urinary tract infection and chronic pyelonephritis-Reflux nephropathy. Clin Nephrol 1:132-141, 1973 19. Smellie J, Edwards D, Hunter N, et al: Vesicoureteric reflux and renal scarring. Kidney lnt 8:S-65-S-72, 1975 (suppl) 20. Ransley PG, Risdon RA: Reflux and renal scarring. BrJ Radio151:lX, 1978 (suppl 14) 21. Edwards D, Normand ICS, Prescod N, et al: Disappearance of vesicoureteric reflux during long-term prophylaxis of urinary tract infection in children. Br Med J 2:285-288, 1977 22. Skoog SJ, Belman AB, Majd M: A nonsurgical approach to the management of primary vesicoureteral reflux. J Urol 138:941-946, 1987 23. Busch R, Huland H: Correlation of symptoms and results of direct bacterial localization in patients with urinary tract infections. J Urol 132:282-285, 1984 24. Sheldon CA, Gonzalez R: Differentiation of upper and lower urinary tract infections: How and when? Med Clin North Am 68:321-323, 1984 25. June CH, Browning MD, Smith LP, et al: Ultrasonography and computed tomography in severe urinary tract infection. Arch Intern Med 145:841-845, 1985 26. Montgomery P, Kuhn JP, Afshani E: CT evaluation of severe renal inflammatory disease in children: Pediatr Radiol 17:216-222, 1987 27. Raynaud C, Tran-Dinh S, Bourguignon M, et al: Acute pyelonephritis in children. Preliminary results obtained with NMR imaging. Contrib Nephrol 56:129-134, 1987 28. Mendez G Jr, Morillo G, Alonso M, et al: Gallium-67 radionuclide imaging in acute pyelonephritis. AIR Am J Roentgenol 134:17-22, 1980 29. Fawcett HD, Goodwin DA, Lantieri RL: I n - l l l leukocyte scanning in inflammatory renal disease. Clin Nucl Med 6:237-241, 1981 30. Handmaker H: Nuclear renal imaging in acute pyelonephritis: Semin Nud Med 12:245-253, 1982 31. Traisman ES, Conway JJ, Traisman HS, et al: The localization of urinary tract infection with 99~Tcglucoheptohate scintigraphy. Pediatr Radiol 16:403-406, 1986 32. Sty JR, Wells RG, Starshak R J, et al: Imaging in

ACUTE PYELONEPHRITIS

acute renal infection in children. A JR Am J Roentgenol 148:471-477, 1987 33. Rushton HG, Majd M, Chandra R, et al: Evaluation of ~mTc-dimercaptosuccicic acid renal scans in experimental acute pyelonephritis in piglets. J Uro1140:1169-1174, 1988 34. Parkhouse HF, Godley ML, Cooper J, et al: Renal imaging with 99~fc-labelled DMSA in the detection of acute pyelonephritis: An experimental study in the pig. Nucl Med Commun 10:63-70, 1989 35. Androulakakis PA, Ransley PG, Risdon RA, et al: Microvascular changes in the early stage of reflux pyelonephritis. An experimental study in the pig kidney. Eur Urol 13:219-223, 1987 36. Majd M, Rushton HG, Chandra RS: Focal intrarenal blood flow changes in experimental acute pyelonephritis in piglets. Radiology 177(P):142, 1990 (suppl, abstr) 37. Conway JJ: The role of scintigraphy in urinary tract infection. Semin Nucl Med 18:308-319, 1988 38. BjOrgvinsson E, Majd M, Eggli KD: Diagnosis of acute pyelonephritis in children: Comparison of sonography and 99mTC-DMSA scintigraphy. AJR Am J Roentgenol 157:539-543, 1991 39. Israel V, Darabi A, McCracken GH: The role of bacterial virulence factors and Tamm-Horsfall protein in the pathogenesis of Escherichia coli urinary tract infection in infants. Am J Dis Child 141:1230-1234, 1987 40. Jodal U: The natural history of bacteriuria in childhood. Infect Dis Clin North Am 1:713-729, 1987 41. Winter AL, Hardy BE, Alton DJ, et al: Acquired renal scars in children. J Urol 129:1190-1194, 1983 42. Smellie JM, Ransley PG, Normand ICS: Development of new renal scars: A collaborative study. Br Med J 290:1957-1960, 1985 43. Lomberg H, Hanson LA, Jacobsson B, et al: Correlation of P blood group, vesicoureteral reflux, and bacterial

111

attachment in patients with recurrent pyelonephritis. N Engl J Med 308:1189-1192, 1983 44. Majd M, Rushton HG, Jantaush B, et al: Relationship among vesicoureteral reflux, P-ftmbriated Escherichia coli, and acute pyelonephritis in children with febrile urinary tract infection. J Pediatr 119:578-585, 1991 45. Rushton HG, Maid M, Jantausch B, et al: Renal scarring following reflux and nonreflux pyelonephritis in children. J Uro147, 1992 (in press) 46. Smellie JM, Normand ICS: Bacteriuria, reflux, and renal scarring. Arch Dis Child 50:581-585, 1975 47. Askari A, Belman AB: Vesicoureteral reflux in black girls. J Urol 127:747-748, 1982 48. BergstrOm T, Larson H, Lincoln K, et al: Studies of urinary tract infections in infancy and childhood. XII. Eighty consecutive patients with neonatal infection. J Pediatr 80:858-866, 1972 49. Svanborg Eden C, Hanson CA, Jodal U, et al: Variable adherence to normal human urinary tract epithelial cells of Escherichia coil strains associated with various forms of urinary infection. Lancet 2:490-492, 1976 50. Lettter H, Svanborg Ed6n C: Chemical identification of glycosphingolipid receptor for Escherichia coil attaching to human urinary tract epithelial cells and agglutinating human erythrocytes. FEMS Microbiol Lett 7:127-134, 1980 51. Khllenius G, M611by R, Svenson SB, et al: The pk antigen as receptor for the hemagglutination pyelonephritis E. coli. FEMS Microbioi Lett 7:297-302, 1980 52. K/illenius G, Svenson SB, Huitberg H, et al: Occurence of P-fimbriated Escherichia coli in urinary tract infections. Lancet 2:1369-1372, 1981 53. Roberts JA, Suarez GM, Haack B, et al: Experimental pyelonephritis in the monkey. VII. Ascending pyelonephritis in the absence of vesicoureteral reflux. J Urol 133:1068-1075, 1985