Kliegman: Nelson Textbook of Pediatrics, 18th ed.

Urinary Tract Infections

PREVALENCE AND ETIOLOGY.

Urinary tract infections (UTIs) occur in 3–5% of girls and 1% of boys. In girls, the first
UTI usually occurs by the age of 5 yr, with peaks during infancy and toilet training. After the
first UTI, 60–80% of girls will develop a second UTI within 18 mo. In boys, most UTIs occur
during the 1st yr of life; UTIs are much more common in uncircumcised boys. The prevalence of
UTIs varies with age. During the 1st yr of life, the male : female ratio is 2.8–5.4 : 1. Beyond 1–2
yr, there is a striking female preponderance, with a male : female ratio of 1 : 10.
UTIs are caused mainly by colonic bacteria. In females, 75–90% of all infections are
caused by Escherichia coli, followed by Klebsiella spp. and Proteus spp. Some series report that
in males older than 1 yr of age, Proteus is as common a cause as E. coli; others report a
preponderance of gram-positive organisms in males. Staphylococcus saprophyticus and
enterococcus are pathogens in both sexes. Viral infections, particularly adenovirus, also may
occur, especially as a cause of cystitis.
UTIs have been considered an important risk factor for the development of renal
insufficiency or end-stage renal disease in children. Some researchers have questioned the
importance of UTI as a risk factor, because only 2% of children with renal insufficiency report a
history of UTI. This paradox may be secondary to better recognition of the risks of UTI and
prompt diagnosis and therapy.

CLINICAL MANIFESTATIONS AND CLASSIFICATION.

The 3 basic forms of UTI are pyelonephritis, cystitis, and asymptomatic bacteriuria.
Clinical pyelonephritis is characterized by any or all of the following: abdominal or
flank pain, fever, malaise, nausea, vomiting, and, occasionally, diarrhea. Newborns may show
nonspecific symptoms such as poor feeding, irritability, and weight loss. Pyelonephritis is the
most common serious bacterial infection in infants <24 mo of age who have fever without a
focus (see Chapter 175 ). These symptoms are an indication that there is bacterial involvement of
the upper urinary tract. Involvement of the renal parenchyma is termed acute pyelonephritis,
whereas if there is no parenchymal involvement, the condition may be termed pyelitis. Acute
pyelonephritis may result in renal injury, termed pyelonephritic scarring.
Acute lobar nephronia (acute lobar nephritis) is a localized renal bacterial infection
involving >1 lobe that represents either a complication of pyelonephritis or an early stage in the
development of a renal abscess. Manifestations are identical to pyelonephritis; renal imaging
demonstrates the abnormality (Fig. 538-1 ).
Renal abscess may occur following a pyelonephritis or may be secondary to a primary
bacteremia (S. aureus).
Perinephric abscesses may be secondary to contiguous infection in the perirenal area
(e.g., vertebral osteomyelitis, psoas abscess) or pyelonephritis that dissects to the renal capsule.
Cystitis indicates that there is bladder involvement; symptoms include dysuria, urgency,
frequency, suprapubic pain, incontinence, and malodorous urine. Cystitis does not cause fever
and does not result in renal injury. Malodorous urine, however, is not specific for a UTI.
Asymptomatic bacteriuria refers to a condition that results in a positive urine culture
without any manifestations of infection. It is most common in girls. The incidence is 1–2% in
preschool and school-age girls and 0.03% in boys. The incidence declines with increasing age.
This condition is benign and does not cause renal injury, except in pregnant women, in whom
asymptomatic bacteriuria, if left untreated, can result in a symptomatic UTI. Some girls are
mistakenly identified as having asymptomatic bacteriuria, whereas they actually are
symptomatic, experiencing day or night incontinence or perineal discomfort.

PATHOGENESIS AND PATHOLOGY.

Virtually all UTIs are ascending infections. The bacteria arise from the fecal flora,
colonize the perineum, and enter the bladder via the urethra. In uncircumcised boys, the bacterial
pathogens arise from the flora beneath the prepuce. In some cases, the bacteria causing cystitis
ascend to the kidney to cause pyelonephritis. Rarely, renal infection may occur by hematogenous
spread, as in endocarditis or in some neonates.
If bacteria ascend from the bladder to the kidney, acute pyelonephritis may occur.
Normally the simple and compound papillae in the kidney have an antireflux mechanism that
prevents urine in the renal pelvis from entering the collecting tubules. However, some compound
papillae, typically in the upper and lower poles of the kidney, allow intrarenal reflux. Infected
urine then stimulates an immunologic and inflammatory response. The result may cause renal
injury and scarring.
Children of any age with a febrile UTI may have acute pyelonephritis and subsequent
renal scarring. The exception to this observation is that if a child with UTIs has a normal 2,3-
dimercaptosuccinic acid (DMSA) scan by 4 yr of age, his or her risk of pyelonephritogenic
scarring from future UTIs is low. Host risk factors for UTI are listed in Table 538-1 .

TABLE 538-1 -- Risk Factors for Urinary Tract Infection

Female gender
Uncircumcised male
Vesicoureteral reflux[*]
Toilet training
Voiding dysfunction
Obstructive uropathy
Urethral instrumentation
Wiping from back to front in females
Bubble bath?
Tight clothing (underwear)
Pinworm infestation
Constipation
Bacteria with P fimbriae
Anatomic abnormality (labial adhesion)
Neuropathic bladder
Sexual activity
Pregnancy
* Risk increased for clinical pyelonephritis, not cystitis.

In girls, UTIs often occur at the onset of toilet training because of voiding dysfunction
that occurs at that age. The child is trying to retain urine to stay dry, yet the bladder may have
uninhibited contractions forcing urine out. The result may be high-pressure, turbulent urine flow
or incomplete bladder emptying, both of which increase the likelihood of bacteriuria. Voiding
dysfunction may occur in the toilet-trained child who voids infrequently. Similar problems may
arise in school-age children who refuse to use the school bathroom. Obstructive uropathy
resulting in hydronephrosis increases the risk of UTI because of urinary stasis. Urethral
instrumentation during a voiding cystourethrogram or nonsterile catheterization may infect the
bladder with a pathogen. Constipation can increase the risk of UTI because it may cause voiding
dysfunction.
The pathogenesis of UTI is based in part on the presence of bacterial pili or fimbriae on
the bacterial surface. There are two types of fimbriae, type I and type II. Type I fimbriae are
found on most strains of E. coli. Because attachment to target cells can be blocked by D-
mannose, these fimbriae are referred to as “mannose-sensitive.” They have no role in
pyelonephritis. The attachment of type II fimbriae is not inhibited by mannose, and these are
known as “mannose-resistant.” These fimbriae are expressed by only certain strains of E. coli.
The receptor for type II fimbriae is a glycosphingolipid that is present on both the uroepithelial
cell membrane and red blood cells. The Gal 1–4 Gal oligosaccharide fraction is the specific
receptor. Because these fimbriae can agglutinate by P blood group erythrocytes, they are known
as P fimbriae. Bacteria with P fimbriae are more likely to cause pyelonephritis. Between 76–94%
of pyelonephritogenic strains of E. coli have P fimbriae, compared with 19–23% of cystitis
strains.
Other host factors for UTI include anatomic abnormalities precluding normal micturition,
such as a labial adhesion. This lesion acts as a barrier and causes vaginal voiding. A neuropathic
bladder may cause UTIs if there is incomplete bladder emptying or detrusor-sphincter
dyssynergia, or both. Sexual activity is associated with UTIs in girls, in part because of
incomplete bladder emptying. From 4–7% of pregnant women have asymptomatic bacteriuria,
which can develop into a symptomatic UTI. The incidence of UTI in infants who are breastfed is
lower than in those fed with formula.
Xanthogranulomatous pyelonephritis is a rare type of renal infection characterized by
granulomatous inflammation with giant cells and foamy histiocytes. It may present clinically as a
renal mass or an acute or chronic infection. Renal calculi, obstruction, and infection with Proteus
spp. or E. coli contribute to the development of this lesion, which usually requires total or partial
nephrectomy.

DIAGNOSIS
A UTI may be suspected based on symptoms or findings on urinalysis, or both, but a
urine culture is necessary for confirmation and appropriate therapy. The correct diagnosis of UTI
depends on having the proper sample of urine. There are several ways to obtain a urine sample;
some are more accurate than others.
In toilet-trained children, a midstream urine sample usually is satisfactory. Most studies
have failed to show any benefit to formally cleansing the introitus before obtaining the specimen.
If the culture shows >100,000 colonies of a single pathogen, or if there are 10,000 colonies and
the child is symptomatic, the child is considered to have a UTI. In uncircumcised males, the
prepuce must be retracted; if the prepuce is not retractable, this method of urine collection may
be unreliable.
In infants, the application of an adhesive, sealed, sterile collection bag after disinfection of
the skin of the genitals can be useful, particularly if the culture is negative. A positive culture
may reflect a contaminant, particularly in girls and uncircumcised boys. In such cases, if the
urinalysis result is positive, the patient is symptomatic, and there is a single organism cultured
with a colony count greater than 100,000, there is a presumed UTI. If any of these criteria are not
met, confirmation of infection with a catheterized sample is recommended.
When greater accuracy regarding the possibility of infection is needed, a catheterized
specimen must be obtained. Proper skin preparation and good catheterization technique are
important. Use of a No. 5 French polyethylene feeding tube in infants or a No. 8 French tube
with proper lubrication in older children minimizes the chance of urethral trauma and
contamination. Only a few milliliters need to be aspirated with a syringe to obtain the urine
sample. Catheterization shortly after spontaneous voiding produces a measure of the residual
urine in the bladder and helps assess problems related to bladder emptying.
Prompt plating of the urine sample is important, because if the urine sits at room
temperature for more than 60 min, overgrowth of a minor contaminant may suggest a UTI when
the urine may not, in fact, be infected. Refrigeration is a reliable method of storing the urine until
it can be cultured.
A urinalysis should be obtained from the same specimen that was cultured. Pyuria
(leukocytes in the urine) suggests infection, but infection can occur in the absence of pyuria;
consequently, this finding is more confirmatory than diagnostic. Conversely, pyuria can be
present without UTI. Nitrites and leukocyte esterase usually are positive in infected urine.
Microscopic hematuria is common in acute cystitis. White blood cell casts in the urinary
sediment suggest renal involvement, but in practice these are rarely seen. If the child is
asymptomatic and the urinalysis result is normal, it is unlikely that there is a UTI. However, if
the child is symptomatic, a UTI is possible, even if the urinalysis result is negative.
With acute renal infection, leukocytosis, neutrophilia, and elevated erythrocyte
sedimentation rate and C-reactive protein are common. The latter two are nonspecific markers of
bacterial infection, and their elevation does not mean that the child has acute pyelonephritis.
With a renal abscess, the white blood cell count is markedly elevated to >20,000–25,000/mm3.
Because sepsis is common in pyelonephritis, particularly in infants and in any child with
obstructive uropathy, blood cultures should be considered.
Acute hemorrhagic cystitis often is caused by E. coli; it also has been attributed to
adenovirus types 11 and 21. Adenovirus cystitis is more common in males; it is self-limiting,
with hematuria lasting approximately 4 days.
Eosinophilic cystitis is a rare form of cystitis of obscure origin that occasionally is found
in children. The usual symptoms are those of cystitis with hematuria, ureteral dilation with
occasional hydronephrosis, and filling defects in the bladder caused by masses that consist
histologically of inflammatory infiltrates with eosinophils. Children with eosinophilic cystitis
may have been exposed to an allergen. Bladder biopsy often is necessary to exclude a neoplastic
process. Treatment usually includes antihistamines and nonsteroidal anti-inflammatory agents,
but in some cases intravesical dimethyl sulfoxide instillation is necessary.
Interstitial cystitis is characterized by irritative voiding symptoms such as urgency,
frequency, and dysuria, and bladder and pelvic pain relieved by voiding with a negative urine
culture. The disorder is most likely to affect adolescent girls and is idiopathic (see Chapter 519.1
). Diagnosis is made by cystoscopic observation of mucosal ulcers with bladder distention.
Treatments have included bladder hydrodistention and laser ablation of ulcerated areas, but no
treatment provides sustained relief.

TREATMENT
Acute cystitis should be treated promptly to prevent possible progression to
pyelonephritis. If the symptoms are severe, a specimen of bladder urine is obtained for culture,
and treatment is started immediately. If the symptoms are mild or the diagnosis is doubtful,
treatment can be delayed until the results of culture are known, and the culture can be repeated if
the results are uncertain. For example, if a midstream culture grows between 104and 105colonies
of a gram-negative organism, a second culture may be obtained by catheterization before
treatment is initiated. If treatment is initiated before the results of a culture and sensitivities are
available, a 3- to 5-day course of therapy with trimethoprim-sulfamethoxazole is effective
against most strains of E. coli. Nitrofurantoin (5–7 mg/kg/24 hr in 3 to 4 divided doses) also is
effective and has the advantage of being active against Klebsiella-Enterobacter organisms.
Amoxicillin (50 mg/kg/24 hr) also is effective as initial treatment but has no clear advantages
over sulfonamides or nitrofurantoin.
 In acute febrile infections suggestive of pyelonephritis, a 10- to 14-day course of broad-
spectrum antibiotics capable of reaching significant tissue levels is preferable. Children who are
dehydrated, are vomiting, or are unable to drink fluids, are ≤1 mo of age, or in whom urosepsis is
a possibility should be admitted to the hospital for intravenous rehydration and intravenous
antibiotic therapy. Parenteral treatment with ceftriaxone (50–75 mg/kg/24 hr, not to exceed 2 g)
or ampicillin (100 mg/kg/24 hr) with an aminoglycoside such as gentamicin (3–5 mg/kg/24 hr in
1 to 3 divided doses) is preferable. The potential ototoxicity and nephrotoxicity of
aminoglycosides should be considered, and serum creatinine and trough gentamicin levels must
be obtained before initiating treatment, as well as daily thereafter as long as treatment continues.
Treatment with aminoglycosides is particularly effective against Pseudomonas spp., and
alkalinization of urine with sodium bicarbonate increases their effectiveness in the urinary tract.
Oral 3rd-generation cephalosporins such as cefixime are as effective as parenteral ceftriaxone
against a variety of gram-negative organisms other than Pseudomonas, and these medications are
considered by some authorities to be the treatment of choice for oral therapy. Nitrofurantoin
should not be used routinely in children with a febrile UTI because it does not achieve
significant renal tissue levels. The oral fluoroquinolone ciprofloxacin is an alternative agent for
resistant microorganisms, particularly Pseudomonas, in patients older than 17 yr. It also has been
used in younger children with cystic fibrosis and pulmonary infection secondary to
Pseudomonas and is used on occasion for short-course therapy in children with Pseudomonas
UTI. However, the clinical use of fluoroquinolones in children should be restricted because of
potential cartilage damage that has been seen in research with immature animals. The safety and
efficacy of oral ciprofloxacin in children is under study. In some children with a febrile UTI,
intramuscular injection of a loading dose of ceftriaxone followed by oral therapy with a 3rd-
generation cephalosporin is effective. A urine culture 1 wk after the termination of treatment of a
UTI ensures that the urine is sterile; in most children, this is unnecessary, because the cultures
often are negative.
Children with a renal or perirenal abscess or with infection in obstructed urinary tracts
often require surgical or percutaneous drainage in addition to antibiotic therapy and other
supportive measures.
In a child with recurrent UTIs, identification of predisposing factors is beneficial. Many
school-aged girls have voiding dysfunction; treatment of this condition often reduces the
likelihood of recurrent UTI. Some children with urinary tract infections void infrequently, and
many also have severe constipation. Counseling of parents and patients to try to establish more
normal patterns of voiding and defecation may be helpful in controlling recurrences. Prophylaxis
against reinfection, using sulfamethoxazole-trimethoprim, trimethoprim, or nitrofurantoin at⅓of
the normal therapeutic dose once a day, often is effective. Prophylaxis with amoxicillin or
cephalexin also may be effective, but the risk of breakthrough UTI may be higher because
bacterial resistance may be induced. Other indications for long-term prophylaxis (e.g.,
neurogenic bladder, urinary tract stasis and obstruction, reflux, calculi) are discussed in other
chapters. There is interest in probiotic therapy, which replaces normal vaginal flora, and
cranberry juice, which prevents bacterial adhesion and biofilm formation, but these agents have
not proved beneficial in preventing UTI.
The main consequences of chronic renal damage caused by pyelonephritis are arterial
hypertension and renal insufficiency; when they are found they should be treated appropriately/

IMAGING STUDIES
The goal of imaging studies in children with a UTI is to identify anatomic abnormalities
that predispose to infection. In children with clinical pyelonephritis (febrile UTI), a renal
sonogram should be obtained to rule out hydronephrosis and structural urinary abnormalities;
sonography also may suggest acute pyelonephritis by demonstrating an enlarged kidney. Power
Doppler sonography has been slightly more sensitive but is unreliable in identifying all cases.
Sonography demonstrates many but not all renal scars. Normally, the difference in renal lengths
between the two kidneys is less than 1 cm, and a larger disparity may be an indication of
impaired renal growth. One should remember that in a child with acute pyelonephritis, a small
kidney may be enlarged because of the infection, giving the erroneous impression that the
kidneys are equal in size. Renal sonography also is sensitive for detecting nephronia and
pyonephrosis, a condition that may require prompt drainage of the collecting system by
percutaneous nephrostomy.
Indications for a voiding cystourethrogram (VCUG) are controversial and are changing.
Most clinicians recommend it for all children with a febrile UTI. A VCUG also is recommended
in girls who have had 2 or 3 UTIs in a period of 6 mo, and for boys with more than one UTI. A
VCUG also should be obtained if the renal sonogram shows any significant abnormality, such as
hydronephrosis, disparity in renal length, or bladder wall thickening. The most common finding
is vesicoureteral reflux, which is identified in approximately 40% of patients ( Fig. 538-3 ).
Timing of the VCUG is controversial. Although in some centers the study is delayed for 2–6 wk
to allow inflammation in the bladder to resolve, the incidence of reflux is identical, regardless of
whether the VCUG is obtained during treatment of the UTI or after 6 wk. Consequently,
obtaining the VCUG before the child is discharged from the hospital is appropriate and ensures
that the evaluation is complete. If available, a radionuclide VCUG rather than a contrast VCUG
can be used in girls; this technique causes less radiation exposure to the gonads than does the
contrast study. However, the radioisotopic VCUG does not provide anatomic definition of the
bladder, allow precise grading of reflux, demonstrate a paraureteral diverticulum, or show
whether reflux is occurring into a duplicated collecting system or an ectopic ureter. In boys,
radiographic definition of the urethra is important; consequently, contrast VCUG is
recommended for the initial work-up.
Because of concern that the VCUG may be traumatic to the child, some parents question
the need for a VCUG if the ultrasonogram is normal. Ultrasonography is insensitive in detecting
reflux; only 40% of children with reflux have any abnormality on the ultrasonogram. The VCUG
should not be performed routinely using general anesthesia, because the study is incomplete
without a voiding phase and it subjects the child unnecessarily to the risk and cost of anesthesia.
In selected cases, oral or nasal midazolam (up to 0.5 mg/kg oral route, 0.2 mg/kg nasal route),
which causes anterograde amnesia and anxiolysis, may be used. Vital signs are monitored and
pulse oximetry is used; no anesthesiologist is present.
When the diagnosis of acute pyelonephritis is uncertain, renal scanning with technetium-
labeled DMSA or glucoheptonate is useful. The presence of photopenia supports the diagnosis of
pyelonephritis, and experienced radiologists can differentiate between an acute and a chronic
process. In approximately 50% of children with a febrile UTI, irrespective of age, the DMSA
scan demonstrates parenchymal involvement. Among children with grade III, IV, or V reflux and
a febrile UTI, 80–90% show acute pyelonephritis. If the DMSA scan shows acute pyelonephritis,
approximately 50% of children will acquire a scar in that site over the following 5 mo. However,
if the DMSA scan is normal during a febrile UTI, no scarring will result from that particular
infection. Computed tomography is another diagnostic tool that can diagnose acute
pyelonephritis, but clinical experience with DMSA is much greater. An alternative and rational
imaging algorithm has been proposed for children with a febrile UTI: renal sonogram and
DMSA renal scan, with a VCUG performed if the DMSA shows acute pyelonephritis. Whether
this approach is optimal awaits the results of prospective trials.
A DMSA scan ( Fig. 538-4 ) often is performed in the presence of vesicoureteral reflux to
assess whether renal scarring is present. The DMSA is the most sensitive and accurate study for
demonstrating scarring. Excretory urography is not as sensitive as the DMSA scan in
demonstrating renal scarring; in addition, visualization of the collecting system in infants and
young children often is suboptimal, there is a slight risk of a contrast allergy, and it can take 1–2
yr for a renal scar to appear on the urogram. CT scanning also has been used to evaluate the
upper urinary tract, because it is effective in demonstrating renal scarring
 In the past, cystoscopy and measurements of urethral caliber often were performed in
girls with UTIs, but these studies contribute nothing to the therapeutic decisions to be made in
children with UTIs and are contraindicated. Narrowing of the female urethra was once postulated
to be a contributing factor in the development of UTIs, but the urethras of girls with recurrent
UTIs have been found not to be narrower than those of girls without infections.