Objective. Experimental evidence suggests that neutrophils and their metabo
lites play an important role in the pathogenesis of pyelonephritis. The aim
of this study was to investigate the diagnostic value of polymorphonuclear
elastase-a(1)-antitrypsin complex (E-a(1)-Pi) for the detection of acute p
yelonephritis in children.
Methods. Eighty-three patients, 29 boys and 54 girls, 25 days to 14 years o
f age, with first-time symptomatic urinary tract infection were prospective
ly studied. Fifty-seven healthy children served as controls. Dimercaptosucc
inic acid (DMSA) scan and voiding cystourethrography were performed in all
patients. Plasma and urinary E-a(1)-Pi, C-reactive protein (CRP), erythrocy
te sedimentation rate (ESR), neutrophil count, urinary N-acetyl-beta-glucos
aminidase (NAG), N-acetyl-beta-glucosaminidase b (NAG b), and creatinine le
vels were measured in all patients on admission and 3 days after the introd
uction of antibiotics. The same markers were also measured in the control s
ubjects.
Results. Planar DMSA scintigraphy demonstrated changes of acute pyelonephri
tis in 30 of 83 children (group A). It was normal in the remaining 53 child
ren (group B). The sex and age distributions were not significantly differe
nt between the 2 groups, as well as between the patients and the control su
bjects (group C). Nineteen of the 53 children with a normal DMSA had body t
emperature greater than or equal to 38 degrees C, whereas all but 4 childre
n with abnormal DMSA had temperature greater than or equal to 38 degrees C.
Therefore, the temperature was significantly different between these 2 gro
ups. The sensitivity and specificity of fever (greater than or equal to 38
degrees C) as an indicator of renal involvement based on isotopic findings
were 86% and 64%, respectively. Given the significant number of the febrile
children with normal DMSA scintiscans, group B was subdivided into B-1 wit
h 19 febrile children (14 boys and 5 girls) and B-2 with 34 children whose
body temperature was below 38 degrees C (8 boys and 26 girls). The sex and
age distribution was significantly different between groups B-1 and B-2. Th
e mean age of group B-1 was .78 years (range: 28 days to 9 years; median: .
25 years; standard deviation: 2.1). All but 1 child in this group were youn
ger than 1 year of age. In contrast, in group B-2, there were only 4 infant
s, the remaining 30 children were older than 2.5 years (mean age: 6 years;
median: 7 years; standard deviation: 3.5; range: 34 days to 12 years). The
mean duration of fever before hospital admission was 2.8 days for group A a
nd 1.8 days for group B-1. This difference was not statistically significan
t. Similarly, body temperature was not significantly different between thes
e 2 groups. The distribution of plasma E-a(1)-Pi values was normal in the c
ontrol subjects. The sensitivity and specificity of plasma E-a(1)-Pi, as an
indicator of renal involvement, were 96% and 50%, respectively, taking the
95th percentile of the reference range as a cutoff value. However, conside
ring as a cutoff value the level of 72 mu g/dL (95th percentile of group B-
2), its sensitivity and specificity were 74% and 86%, respectively. Plasma
E-a(1)-Pi levels were significantly elevated in group A compared with group
B and in both groups, the plasma E-a(1)-Pi values were significantly highe
r than in the control subjects. A significant difference also was noticed b
etween group A and each of the subgroups B-1 and B-2 and also between the s
ubgroups themselves. Plasma E-a(1)-Pi concentrations correlated significant
ly with neutrophil count in groups A (r = .3), B (r = .4), and B-2 (r = .46
), but the correlation was not significant in group B-1. ESR levels showed,
among the different groups, similar differences with those of E-a(1)-Pi va
lues. Unlike E-a(1)-Pi, CRP levels were comparable between groups A and B-1
, which both consisted of febrile children. Neutrophil count was not signif
icantly different between subgroups B-1 and B-2. Considering 20 mg/dL as a
cutoff level for CRP, its sensitivity and specificity for identifying the u
rinary tract infection site were 69% and 57%, respectively. The sensitivity
and specificity of ESR, using 30 mm/hour as a cutoff value, were 90% and 5
9%, respectively. The comparison of febrile infants with a normal DMSA scan
(all but 1 child of group B-1) with those with an abnormal one (a subpopul
ation of group A) showed significant difference of plasma E-a(1)-Pi and ESR
but not of CRP and neutrophils.
Urinary E-a(1)-Pi, as well as NAG and NAG b/creatinine values, showed no si
gnificant difference between groups A and B. NAG and NAG b levels were sign
ificantly higher in group B-1 compared with group B-2 but they were similar
with those of group A. Reflux was noticed in 16/83 children (19%), 9/30 ch
ildren with an abnormal DMSA (30%) and 7/53 with a normal DMSA scan (13%);
this difference was not statistically significant. The sensitivity and spec
ificity of reflux, as an indicator for renal lesions on the DMSA scan, were
30% and 86%, respectively. The follow-up investigation on the third day re
vealed that plasma E-a(1)-Pi levels, as well as CRP, were significantly low
er compared with their levels on admission within each group. Despite the f
act that ESR levels were lower on the third day, the difference was not sig
nificant.
Conclusions. Plasma E-a(1)-Pi is a sensitive but not a specific marker for
the detection of acute pyelonephritis. Urinary E-a(1)-Pi levels cannot be u
sed for this purpose.