Relationship among vesicoureteral reflux, P-fimbriated Escherichia coil, and acute pyelonephritis in children with febrile urinary tract infection M a s s o u d M a j d , MD, H. Gil Rushton, MD, Barbara J a n t a u s c h , MD, a n d Bernhard L. W i e d e r m a n n , MD From the Departments of Radiology (Nuclear Medicine), Urology, and Infectious Diseases, Children's National Medical Center, and the Departments of Radiology, Urology and Pediatrics, George Washington University School of Medicine, Washington, D.C. Ninety4our children with febrile urinary tract infection were studied prospectively to determine the relationship b e t w e e n vesicoureteral reflux, P-fimbriated Escherichia colt, and a c u t e pyelonephritis, and to e v a l u a t e the diagnostic reliability of c o m m o n l y used clinical and laboratory observations. By using renal scan with d i m e r c a p t o s u c c i n i c acid l a b e l e d with technetium 99m as the standard of reference, we d o c u m e n t e d 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 pyelonephritis. Of the 70 E. c a l l urinary isolates, 48 (69%) were P-fimbriated, including 30 (64%) of 47 isolates from patients with pyelonephritis and 18 (78%) of 23 isolates from patients with normal renal scans. The p r e v a l e n c e of P-fimbriated E. c o l t in patients with pyeionephritis and vesicoureteral reflux was 46%, c o m p a r e d with 71% in those with pyelonephritis who had no concurrent vesicoureteral reflux (p = 0.222). Multiple clinical and laboratory variables c o m m o n l y used in the diagnosis of a c u t e pyelonephritis did not a d e q u a t e l y predict the presence or a b s e n c e of p a r e n c h y m a l involvement. These data show the following: (I) Acute pyelonephritis in the absence of demonstrable vesicoureteral reflux is common. (2) Febrile urinary tract infections in children are c o m m o n l y associated with P-fimbriated E. colt, both in the presence and a b s e n c e of vesicoureteral reflux. (3) The presence of P fimbriae a l o n e does not tully explain the p a t h o p h y s i o l o g y of renal parenchymal invasion by bacteria in the a b s e n c e of vesicoureteral reflux. (4) The diagnosis of acute pyelonephritis in children with febrile urinary tract infections on the basis of clinical and laboratory observations is unreliable. (J PEDIATR
1991;119:578-85) Recent reports have challenged the assumptions that acute pyelonephritis in children is usually associated with vesicoureteral reflux, k 2 Acute pyelonephritis in the absence of Submitted for publication Sept. 21, 1990; accepted May 7, 1991, Reprint requests: Massoud Maid, MD, Department of Diagnostic Imaging & Radiology, Children's Ngtional Medical Center, 111 Michigan Ave., N.W., Washington, DC 20010. 9/2o/3o8os
578
vesicoureteral reflux has been attributed to bacterial adherence to uroepithelial cells related to the presence of P fimbriae in certain strains of Escherichia coll. ~ The differentiation of acute pyelonephritis from lower urinary tract infections on the basis of classic clinical and laboratory findings may be difficult, especially in children. 4-6 In an experimental animal study, we found renal cortical scintigraphy with dimercaptosuccinic acid labeled with technetium 99m to be highly sensitive and reliable for
Volume I 19 Number 4
CFU CVA DMSA ESR UTI
Vesicoureteral reflux and febrile UTI
Colony-forming units Costovertebra[ angle Dimercaptosuccinic acid Erythrocyte sedimentation rate (Westergren) Urinary tract infection
detecting and localizing early parenchymal inflammatory changes. 7 It therefore seemed reasonable to use 99mTcDMSA scans clinically to confirm the diagnosis of acute pyelonephritis. The aims of this prospective clinical study were as follows: (1) to determine the prevalence of acute pyelonephritis in the absence of vesicoureteral reflux in children with febrile UTI, (2) to investigate the significance of P-fimbriated E. eoli in the pathogenesis of acute pyelonephritis in the absence of reflux, and (3) to determine the predictive accuracy of commonly used clinical and laboratory findings for diagnosing acute pyelonephritis, with the 99mTc-DMSA renal scan used as the standard of reference. METHODS
The following criteria were chosen for inclusion of patients in the study: all patients who were hospitalized with a diagnosis of febrile UTI and acute pyelonephritis were eligible, provided that body temperature was at least 38 ~ C (100.5 ~ F) within 24 hours of admission and urine culture grew a single organism with a colony count of at least 105 C F U / m l on a clean-catch specimen or at least 104 C F U / ml on a catheterized specimen. Ninety-four of 116 consecutive patients, from March 1987 to December 1988, met the criteria. Of the 94 patients, 16 (17%) had underlying conditions that could have predisposed them toward UTI (spina bifida, 10; cerebral palsy with atonic bladder, one; diabetes, one; posterior urethral valve, two; ureteropelvic junction obstruction, one; and ectopic ureter, one). Clinical signs and symptoms included maximum daily temperature, duration of fever, and the presence of chills, lethargy, nausea, vomiting, anorexia, dysuria, abnormal urinary frequency, enuresis, costovertebral angle tenderness, and abdominal tenderness. Laboratory studies included complete blood cell count, determination of the Westergren crythrocyte sedimentation rate and serum creatinine level, urinalysis, urine culture, and blood culture. In addition, urinary isolates from the patients with E. eoli infection were studied for the presence of P fimbriae, as determined by mannose-resistant hemagglutination of OP1 but not Op erythrocytes. 8 Renal ultrasonography and 99mTc-DMSA renal cortical scintigraphy were carried out within 72 hours of admission. Cystography was performed in all patients before their discharge from the hospital. The sonograms were obtained by
579
using a real-time scanner equipped with 3.5 and 5 MHz transducers. The kidneys were evaluated for size, shape, echogenicity, cortical scars, hydronephrosis, and o t h e r structural abnormalities. The renal cortical scans were obtained 2 to 3 hours after an intravenous injection of 99mTc-DMSA in a dose of 50 #Ci/kg (minimum 300 ~Ci). High-resolution magnified images of the kidneys in posterior and posterior-oblique projections were obtained routinely by using a gamma camera-computer system equipped with a pinhole collimator. In addition to general evaluation for renal size, shape, location, and differential function, the images were evaluated for evidence of acute pyelonephritis and cortical scarring. The criterion for acute pyelonephritis was the presence of areas of focal (single or multiple) or diffuse decreased cortical uptake of 99mTc-DMSA without evidence of cortical loss (Fig. 1). Defects in uptake associated with cortical thinning and decreased volume were considered chronic scars (Fig. 2). In our experience, this criterion is reliable in differefitiating acute lesions from cortical scars except in severely scarred kidneys, in which the diagnosis of an acute lesion adjacent to a scar may be difficult. The cystograms were obtained with either radiographic technique (voiding cystourethrography with iodinated contrast medium) or with direct radionuclide technique (isotope cystography with 99mTc pertechnetate), The radiographic cystograms were evaluated for the presence and grade of reflux by using the International Grading System. 9 The reflux demonstrated by radionuelide technique was graded as mild, moderate, or severe. For purposes of comparison, radiographic grades I and I1 were classified as mild, grades III and IV were classified as moderate, and grade V was classified as severe. Statistical analysis consisted of two phases. The first phase included bivariable analyses, in which characteristics of patients in two groups were compared. When statistical hypotheses were tested, continuous variables were dichotomized and the chi-square procedure corrected for continuity was used (a = 0.05). When the assumptions of that procedure were suspect, the Fisher Exact Test was substituted. The second phase of statistical analysis examined the diagnostic accuracy of various measurements, with control for differences in age, gender, and race. This was accomplished by using logistic regression analysis, with acute pyelonephritis determined by scan as the dependent variable. To control for age, we investigated two approaches. First, age groups (birth through 1, 1 through 3, 3 through 6, and greater than 6 years) were represented by indicator variables in the regression equation. Second, age was considered a continuous independent variable; in this approach, polynomial functions of age were examined to allow a curvilin-
580
M a j d et al.
The Journal of Pediatrics October 1991
L
"
A
R
C
B
Fig. I. Different scintigraphic patterns of acute pyelonephritis. A, Focal pyelonephritis--decreasedtracer uptake in upper poIe of right kidney (arrow). B, Multifoca/pyelonephritis--multiplediscrete areas of decreased tracer uptake in swollen left kidney. C, Diffuse pyelonephritis--swollen right kidney with diffusely decreased tracer uptake. ear relationship. This second approach was considered the better of the two, because age and the square of age contributed more to diagnostic accuracy than did age group. Determination of diagnostic accuracy from logistic regression analysis was performed separately for each characteristic examined. Characteristics representing continuous data were dichotomized for this analysis. To begin, a regression equation including age, the square of age, gender, race (white, black, or other), an indicator of the characteristic, and all two-way interactions was obtained. Then, by backward elimination using the criterion o f p __
1991;119:578-85) Recent reports have challenged the assumptions that acute pyelonephritis in children is usually associated with vesicoureteral reflux, k 2 Acute pyelonephritis in the absence of Submitted for publication Sept. 21, 1990; accepted May 7, 1991, Reprint requests: Massoud Maid, MD, Department of Diagnostic Imaging & Radiology, Children's Ngtional Medical Center, 111 Michigan Ave., N.W., Washington, DC 20010. 9/2o/3o8os
578
vesicoureteral reflux has been attributed to bacterial adherence to uroepithelial cells related to the presence of P fimbriae in certain strains of Escherichia coll. ~ The differentiation of acute pyelonephritis from lower urinary tract infections on the basis of classic clinical and laboratory findings may be difficult, especially in children. 4-6 In an experimental animal study, we found renal cortical scintigraphy with dimercaptosuccinic acid labeled with technetium 99m to be highly sensitive and reliable for
Volume I 19 Number 4
CFU CVA DMSA ESR UTI
Vesicoureteral reflux and febrile UTI
Colony-forming units Costovertebra[ angle Dimercaptosuccinic acid Erythrocyte sedimentation rate (Westergren) Urinary tract infection
detecting and localizing early parenchymal inflammatory changes. 7 It therefore seemed reasonable to use 99mTcDMSA scans clinically to confirm the diagnosis of acute pyelonephritis. The aims of this prospective clinical study were as follows: (1) to determine the prevalence of acute pyelonephritis in the absence of vesicoureteral reflux in children with febrile UTI, (2) to investigate the significance of P-fimbriated E. eoli in the pathogenesis of acute pyelonephritis in the absence of reflux, and (3) to determine the predictive accuracy of commonly used clinical and laboratory findings for diagnosing acute pyelonephritis, with the 99mTc-DMSA renal scan used as the standard of reference. METHODS
The following criteria were chosen for inclusion of patients in the study: all patients who were hospitalized with a diagnosis of febrile UTI and acute pyelonephritis were eligible, provided that body temperature was at least 38 ~ C (100.5 ~ F) within 24 hours of admission and urine culture grew a single organism with a colony count of at least 105 C F U / m l on a clean-catch specimen or at least 104 C F U / ml on a catheterized specimen. Ninety-four of 116 consecutive patients, from March 1987 to December 1988, met the criteria. Of the 94 patients, 16 (17%) had underlying conditions that could have predisposed them toward UTI (spina bifida, 10; cerebral palsy with atonic bladder, one; diabetes, one; posterior urethral valve, two; ureteropelvic junction obstruction, one; and ectopic ureter, one). Clinical signs and symptoms included maximum daily temperature, duration of fever, and the presence of chills, lethargy, nausea, vomiting, anorexia, dysuria, abnormal urinary frequency, enuresis, costovertebral angle tenderness, and abdominal tenderness. Laboratory studies included complete blood cell count, determination of the Westergren crythrocyte sedimentation rate and serum creatinine level, urinalysis, urine culture, and blood culture. In addition, urinary isolates from the patients with E. eoli infection were studied for the presence of P fimbriae, as determined by mannose-resistant hemagglutination of OP1 but not Op erythrocytes. 8 Renal ultrasonography and 99mTc-DMSA renal cortical scintigraphy were carried out within 72 hours of admission. Cystography was performed in all patients before their discharge from the hospital. The sonograms were obtained by
579
using a real-time scanner equipped with 3.5 and 5 MHz transducers. The kidneys were evaluated for size, shape, echogenicity, cortical scars, hydronephrosis, and o t h e r structural abnormalities. The renal cortical scans were obtained 2 to 3 hours after an intravenous injection of 99mTc-DMSA in a dose of 50 #Ci/kg (minimum 300 ~Ci). High-resolution magnified images of the kidneys in posterior and posterior-oblique projections were obtained routinely by using a gamma camera-computer system equipped with a pinhole collimator. In addition to general evaluation for renal size, shape, location, and differential function, the images were evaluated for evidence of acute pyelonephritis and cortical scarring. The criterion for acute pyelonephritis was the presence of areas of focal (single or multiple) or diffuse decreased cortical uptake of 99mTc-DMSA without evidence of cortical loss (Fig. 1). Defects in uptake associated with cortical thinning and decreased volume were considered chronic scars (Fig. 2). In our experience, this criterion is reliable in differefitiating acute lesions from cortical scars except in severely scarred kidneys, in which the diagnosis of an acute lesion adjacent to a scar may be difficult. The cystograms were obtained with either radiographic technique (voiding cystourethrography with iodinated contrast medium) or with direct radionuclide technique (isotope cystography with 99mTc pertechnetate), The radiographic cystograms were evaluated for the presence and grade of reflux by using the International Grading System. 9 The reflux demonstrated by radionuelide technique was graded as mild, moderate, or severe. For purposes of comparison, radiographic grades I and I1 were classified as mild, grades III and IV were classified as moderate, and grade V was classified as severe. Statistical analysis consisted of two phases. The first phase included bivariable analyses, in which characteristics of patients in two groups were compared. When statistical hypotheses were tested, continuous variables were dichotomized and the chi-square procedure corrected for continuity was used (a = 0.05). When the assumptions of that procedure were suspect, the Fisher Exact Test was substituted. The second phase of statistical analysis examined the diagnostic accuracy of various measurements, with control for differences in age, gender, and race. This was accomplished by using logistic regression analysis, with acute pyelonephritis determined by scan as the dependent variable. To control for age, we investigated two approaches. First, age groups (birth through 1, 1 through 3, 3 through 6, and greater than 6 years) were represented by indicator variables in the regression equation. Second, age was considered a continuous independent variable; in this approach, polynomial functions of age were examined to allow a curvilin-
580
M a j d et al.
The Journal of Pediatrics October 1991
L
"
A
R
C
B
Fig. I. Different scintigraphic patterns of acute pyelonephritis. A, Focal pyelonephritis--decreasedtracer uptake in upper poIe of right kidney (arrow). B, Multifoca/pyelonephritis--multiplediscrete areas of decreased tracer uptake in swollen left kidney. C, Diffuse pyelonephritis--swollen right kidney with diffusely decreased tracer uptake. ear relationship. This second approach was considered the better of the two, because age and the square of age contributed more to diagnostic accuracy than did age group. Determination of diagnostic accuracy from logistic regression analysis was performed separately for each characteristic examined. Characteristics representing continuous data were dichotomized for this analysis. To begin, a regression equation including age, the square of age, gender, race (white, black, or other), an indicator of the characteristic, and all two-way interactions was obtained. Then, by backward elimination using the criterion o f p __
