APN has been estimated to have an incidence of 250,000 cases/year in the U.S. [4], and the direct and indirect cost amounts to about 2,140,000,000 US dollars per year [5].

Women are infected five times more often than men, even though they have a lower mortality rate (7.3 vs. 16.5 deaths/1000 cases [4]).

APN is common among diabetics with an incidence of 51.4 and 147.9/1000 person-years for men and women, respectively [6].

The incidence of APN during pregnancy was 0.5% in a series of 543,430 patients [7].

Renal and perirenal abscesses, which may be either of ascending or hematogenic origin, may complicate an infection of the renal parenchyma [8, 9]. Although the problem of renal abscesses is underestimated in the literature, their frequency is high: in the U.S. they are responsible for 1–10/10,000 hospitalizations yearly with a mortality of 0.7–1.6% [9].

The spectrum of APN manifestations ranges from mild symptoms to a full septic syndrome.

The clinical presentation is traditionally a triad, including loin pain, fever, and bacteriuria and/or pyuria [10]; however, it has been demonstrated that bacteriuria and pyuria may not be present, even in cases of APN which have been confirmed by imaging [11–13].

Loin pain is present in 86% of cases, fever in 77%, and at least one of these symptoms in 95% of cases [14], but one-third of elderly patients present without fever and in 20% of cases symptoms are gastrointestinal or pulmonary [15]. Symptoms that are typical of cystitis may accompany or precede the onset of APN in 83% of cases.

Acute renal failure rarely occurs [16], and may be due to dehydration, septic shock, tubular toxicity by gram-negative microorganisms, tubular injury due to diffuse interstitial infiltration of polymorphonuclear cells and bacteria. It is always present when the infection has spread bilaterally throughout the entire renal parenchyma.

Although in most cases the clinical features are typical and allow the diagnosis to be made, the literature indicates the need for microbiological positive finding.

Sobel and Kaye [10] for the Infectious Disease Society of America (IDSA) defined APN as the pathology in which a growth of at least 10,000 cfu/ml is found in the urine in the presence of typical symptoms. Lower levels of bacterial growth (1000–10,000 cfu/ml) may be significant in pregnant women and in men.

Other diagnostic criteria include isolation of the same microorganism in the urine and in the blood [10, 17], or the concomitant presence of loin pain, fever (axillary >38 °C), pyuria (>10³ WBC/high microscopic field—hmf), and positive urine culture (>10⁵ cfu/ml).

However, a review of the literature suggests that this definition is outdated: in fact Gupta and Hooton [12, 13] state that pyuria and bacteriuria may be absent if the infection is not communicating with the urinary tract, or if an obstruction is present.

In a study performed by our group involving 223 patients (202 women, mean age 37.77 ± 17.61 years) tested at their arrival in the Emergency Unit of our hospital, urine culture was positive in only 30% of patients and pyuria in 65% [11]. Moreover, among 196 patients who underwent CT/NMR, only 46 (23.4%) had a positive urine culture. In 98 patients who had positive CT/NMR urine or blood cultures were negative [11].

The low frequency of positive urine cultures may be explained by previous antibiotic treatment, either self-prescribed or prescribed by the general practitioner, and by the possibility that the infection is confined to the renal parenchyma. Moreover, atypical organisms, such as Ureaplasma urealyticum (responsible for 4.8% of APN cases [9]) and Mycoplasma hominis, are not found unless particular culture media containing arginine and urea are used.

Obtaining an antibiogram is very important in order to give the patient a targeted antibiotic. Therefore, we recommend collecting urine as soon as possible before starting antibiotics. However, we believe that a urine culture is not essential nowadays in order to diagnose APN. The same considerations hold true for pyuria.

Blood cultures are positive in about 20% of cases [15] (21.4% of cases in our series [11]), and there is no evidence that they indicate a more severe form of APN [18], or that they should modify the therapeutic strategy [19, 20].

Blood cultures are indicated in case of diagnostic doubt, in a situation of immunosuppression or when a hematogenic source is suspected [19].

Leukocytosis was present in 82.6% of cases and elevated C-reactive protein (CRP) in 100% of cases in our series. Mean CRP was 15.65 ± 8.56 mg/dl [11].

These two parameters, which are commonly tested in clinical practice, may be helpful together with the presence of fever to distinguish a renal colic from an APN.

US is the first investigation that must be performed in order to exclude hydronephrosis, stones, or urinary abnormalities.

US examination can show ureteral thickening, irregular, and focal parenchymal echogenicity (usually hyperechogenicity), increase in kidney size or hypotonia of the intra-renal cavities (Figs. 7.1 and 7.2a, b), while it only rarely evidences abscesses.

Several authors [11, 21–24] envisage that it is necessary to document APN with CT or NMR. These examinations allow a better definition of the extent of the inflammatory lesions, and confirm the diagnosis in case of clinical doubt (atypical pain or negative urine culture). Moreover, CT and NMR are more sensitive at detecting intra- and perirenal abscesses than US [25].

CT must be performed with contrast medium. It shows triangular areas of hypodense parenchyma, the apex being toward the papilla and the base toward the renal cortex. These areas may be multiple and bilateral (Figs. 7.3, 7.4, 7.5 and 7.6a, b).

In APN, no hypodense areas can be seen in the arterial phase, as they only become evident in the nephrographic phase and are better demonstrated in the venous phase (Fig. 7.7a, b). Sometimes they appear as hyperdense lesions in a late phase of the test, 2 h after the injection of the contrast medium.

In renal infarction, an avascular area of renal parenchyma can be seen in the arterial and venous phases; the lesion has sharper margins than APN areas.

The attenuation of density depends on the focal reduction of perfusion due to vessel compression by edema, intravascular granulocyte aggregation, and defective tubular function with altered contrast medium tubular transport and concentration.

NMR has a sensitivity and a specificity similar to that of CT and it is therefore preferable in young women (Figs. 7.8 and 7.9).

«Diffusion NMR» is useful in case of renal failure because it allows good imaging even without contrast medium (Fig. 7.10).

In our study, US was normal in 109 cases (52.1%). CT scan was performed in 183/223 (82.06%) patients and showed lesions suggestive of APN in 170 cases (92.8%), with evidence of single or multiple areas of parenchymal hypodensity. Concordance between CT and US was 49% [11].

NMR was performed in 57 cases (47 positive, 10 negative).

Altogether, 213 patients underwent CT and/or NMR (95.5%) with radiologic confirmation of APN in 196/223 patients (87.9%). Among these patients, only 46 (23.5%) had positive urine cultures, 31 (15.3%) had positive blood cultures, and 15 (7.6%) had positive cultures of both urine and blood. Urine or blood cultures were negative in 98 patients in whom CT/NMR was positive for APN. In 12 patients CT was normal while blood or urine cultures were positive [11] (Table 7.3).

No differences were found between patients with positive or negative CT/NMR as regards body temperature at admission, leukocytosis, CRP, or duration of symptoms before hospitalization.

Similar data were reported in another study and no differences were found between patients with single or multiple inflammatory lesions [25].

We found single or multiple abscesses in 23.5% of patients who underwent CT/NMR, which were evident at US examination in only 2 patients.

We believe that NMR (especially in young women) or CT should be routinely performed in patients with APN since evidence of an abscess would influence the following therapeutic strategy.

Differential diagnosis must be made with several other situations, the most similar to APN is renal infarction, which may present similar clinical manifestations at onset, including fever [26]. CT may distinguish the two disorders, even though the radiologic features can sometimes be difficult to interpret [27] (Figs. 7.11a, b and 7.12).

Pelvic inflammatory disease, cholecystitis, appendicitis, lower lobe pneumonia, ovary or uterine torsion, abdominal abscesses, ovarian cysts, intestinal perforation, and Herpes zoster prodromes may also mimic APN.

Asymptomatic bacteriuria occurs in 2–10% of pregnancies and, if untreated, up to 30% of these patients may develop APN during pregnancy [28].

APN is especially dangerous in pregnancy; therefore, urine cultures must be monitored regularly during pregnancy, and in case of previous APN they must be repeated every week.

APN is more frequent in the second trimester of pregnancy [7]. It seems to be more frequent in nulliparous and in young women [7], and may lead to acute renal dysfunction and spontaneous preterm delivery. Most of the preterm births in the study by Wing et al. occurred between 33 and 36 weeks of gestation (9.1%) [7].

Relapses occurred in 25% of cases [7].

The most frequent etiologic agents are Escherichia coli (70%) and gram-positive microorganisms, especially beta-hemolytic Streptococci (10%).

Diabetes is a common predisposing factor for UTIs, entailing a relative risk of 1.2–2.2 [29]. APN in diabetic patients is 5–10 times more frequent than in nondiabetic patients of both genders [30].

The reason for this is not clear. Geerlings [31] reports increased vaginal colonization by E. coli in diabetic patients, perhaps due to greater bacteria adherence to the cells of the uroepithelium or to an altered or delayed inflammatory response and cytokine secretion.

Klebsiella, Enterobacter, Clostridium, or Candida are the microorganisms that are most often responsible for APN in diabetics.

Kumar et al. [32] evaluated 105 cases of APN in diabetic patients: 24.8% had emphysematous pyelonephritis (EP), 3.8% had renal papillary necrosis, 12.3% had renal abscesses, 39% had bacteremia, and 17% had renal failure.

The outcome in diabetic patients can be poor, and an 11% mortality rate has been reported [32].

Careful attention must be paid to EP [15]. This is a necrotizing infection with gas formation in the renal parenchyma, collecting system or perinephric tissue that develops almost exclusively in diabetic patients. Therefore, it is necessary to maintain a high degree of suspicion and to perform imaging studies early during the course of APN in diabetic patients.

Hyperglycemia has been postulated to be an important factor for gas formation, which requires anaerobic metabolism of glucose [33]. In fact, in the series of diabetic patients reported by Kumar [32], 24.8% with EP, patients with EP had poorer sugar control than patients without EP.

The reported sensitivities of plain X-ray, US, and CT scan for detecting EP are 65, 69, and 100%, respectively [34].

EP may be extremely severe, so that nephrectomy rates in these patients were higher than in non-emphysematous pyelonephritis patients (P < 0.05), and mortality was 30.8% [32].

Renal and perirenal abscesses may complicate an infection of the renal parenchyma, and they may be either of ascending or hematogenic origin [8, 9].

Risk factors for developing abscesses include urinary tract obstruction, multiresistant pathogens, diabetes, recurrent UTIs (66% of cases), stones (30%), instrumentation of the urinary tract, association with VUR, neurogenic bladder, cancer-causing urinary obstruction, simple cysts, and renal polycystic disease [35].

Abscesses may be cortical (75% of cases are observed in men) and cortico-medullar (with similar distribution in genders).

In our series, 50/213 patients studied by CT/NMR had abscesses (23.5%) [11], with only two abscesses being detected by US.

CT shows a defect of perfusion during the arterial phase due to the occlusion of the small vessels by inflammatory cells and edema. During the venous phase, the capsule of the abscess may be evident, becoming hyperdense. In a late phase a lack of concentration of the contrast medium can be observed (Figs. 7.13 and 7.14).

CT without contrast medium detects only large abscesses.

At NMR, the image in T2 is hyperintense in the acute phase because of edema, and hypo-intense in a later phase. After gadolinium injection, hypoperfusion in the arterial phase, delimitation of the rim in the venous phase, and lack of contrast absorption in a late phase are observed (Figs. 7.15 and 7.16).

Another useful examination is «diffusion NMR», which requires only a few minutes when added to a routine NMR, and allows good imaging without contrast medium [36]. Diffusion NMR is a type of functional imaging, where the image contrast is derived from differences in the Brownian motion of water molecules in tissues [37]. Since the signal is derived from the inherent tissue contrast, the administration of intravenous contrast medium is not required. The imaging signal confers information about the biophysical properties of tissues, such as cell organization and density.

In a study on 42 patients with APN, diffusion NMR showed a higher sensitivity (95.3%) than that of contrast-enhanced CT (88.1%) [38].

In another study, agreement between CT and diffusion NMR was 94.3% [39].

Diffusion-weighted imaging appears to be reliable in the diagnosis and follow up of APN and could provide a reasonable alternative to contrast-enhanced magnetic resonance imaging [39].

In our series of patients no differences were found between patients with or without abscesses as regards body temperature (39.3 ± 0.66 vs. 39.16 ± 0.81 °C), leukocytosis (16,912.72 ± 6676.36 vs. 14,979.67 ± 6434/mm³), duration of fever (5.48 ± 4.23 vs. 5.44 ± 7.52 days), duration of symptoms before hospitalization, CRP, pyuria, and urine culture positivity (20 vs. 31.5%). Patients with abscesses were hospitalized for longer periods of time [11] (Table 7.4).

Mortality of patients with abscesses is reported to vary from 1.5 to 15% [9], while in uncomplicated APN mortality is reported to be 0.7–1.6% (4).

In APN at macroscopic examination the kidney appears enlarged and edematous. The surface may look variegated by the presence of yellowish areas of different extensions, which represent parenchymal abscesses.

When the kidney is cut, abscesses localize prevalently in the cortex but exudation may extend to the collecting ducts. Sometimes the lesions attain the perirenal connective tissue.

The distribution of the lesions may be casual, even though the renal poles are more frequently involved.

The renal pelvis is usually dilated; its mucosa may be edematous, reddish, or covered by pus.

In most severe cases the renal papillae are ulcerated or necrotic (papillary necrosis).

The histologic aspect is dominated, at least in the initial phases of the disease, by neutrophil infiltration, which may be diffused or organized in abscesses. Neutrophil infiltration localizes in the interstitium, in tubules and in the interior of the collecting ducts, which appear dilated and, in severe forms, necrotic (Fig. 7.17). Neutrophils collecting in the tubules can form casts (pus casts), which may be found in the urine. Leukocytes sometimes organize to form a granuloma.

Zones with inflammatory infiltrates alternate with spared areas.

Interstitial edema is an early feature, and usually occurs in concert with cellular infiltration [40].

In the early phases, glomeruli and vessels are usually spared, except in hematogenous APN, in which medullary involvement is milder and less frequent.

An infiltration by lymphocytes, plasma cells, and rarely by eosinophils may be associated and become prevalent in late phases.

Pyelocalyceal mucosa is always involved in the inflammatory process in ascending APN where it appears swollen. Infiltrates by granulocytes can be present also on the intraluminal surface.

Special dyes may be useful or even essential to establish the bacterial (Gram) or the fungal etiology (PAS, Silver Metenamine, Grocott).

In chronic pyelonephritis lymphocytes and plasma cells are seen together with interstitial fibrosis and tubular atrophy. Collapsed tubules combine with dilated tubules.

The size of the affected kidney may be normal or slightly enlarged.

In the ascending pathway of infection, the pyelocalyceal mucosa presents purulent exudate. The inflammation and the abscesses have a radial distribution, from the calyces to the renal cortex.

In the hematogenous infection, multiple, isolated pale-yellowish abscesses (1–5 mm in diameter) with a hyperemic halo can be seen on the cut surface, especially in the cortical area and on the surface of the kidney.

Microabscesses may merge and create large purulent cavities.

E. coli represents by far the prevalent etiologic agent, and is present in 56.4% of cases. Enterococci are found in 10.7% of cases, Staphylococcus species in 8%, Proteus mirabilis in 6%, Enterobacter species in 5.3%, and Pseudomonas aeruginosa in 5.3% of cases [17].

In our series [11], we detected E. coli in 56 patients of the 64 with positive urine culture(87.5%). Other pathogens included Klebsiella pneumoniae, Enterococcus faecalis, Proteus mirabilis, and Klebsiella pneumoniae plus Enterococcus faecalis.

Blood cultures were positive in 39/182 patients (21.4%): E. coli in 35 patients, Acinetobacter lwoffii in 1 patient, Proteus mirabilis in 1 patient, Streptococcus saprophyticus in 1 patient, and Staphylococcus hominis in 1 patient [11].

E. coli is less frequently found in elderly patients because, bladder catheters and the frequent instrumentation allow the selection of different microorganisms, above all gram-negative bacteria such as Proteus, Pseudomonas, Klebsiella, and Serratia [4].

The etiologic agents of cortico-medullary abscesses include E. coli (75%), Klebsiella, Proteus, Enterobacter, Serratia (15–20%), Streptococcus faecalis, and Staphylococcus aureus (5%) [9].

Cortical abscesses are prevalently due to S. aureus (90% of cases) [9].

Bacteria usually ascend through the urethra, bladder and ureters to the kidneys.

A healthy bladder can eliminate the microorganisms which have been introduced within two to three days. The defense mechanisms of the bladder include voiding, antibacterial properties of urine, and intrinsic defense mechanisms of the mucosa. Moreover, the acid milieu of the vaginal environment in women and of prostate secretions in men contributes to protect against infections.

Hematogenic infections are more frequent in frail patients and in those affected with chronic disease, or who are immunosuppressed.

Staphylococci and fungi may reach the kidney through the blood stream from distant foci, which may be anywhere, though they are most often found in the skin or bones [4].

Sexual activity in women plays a fundamental role in the pathogenesis of APN. A correlation has been found between APN and frequency, promiscuity, and characteristics of sexual intercourse [14].

The use of spermicides, previous urinary infections, familial (maternal) history of UTIs, smoking habit, and difficulty in holding urine are also significant risk factors for the development of APN [14].

In men, prostatitis and prostate hypertrophy predispose to APN [4].

Other predisposing conditions include urinary tract obstruction, diabetes, immunosuppression, urinary instrumentation, and use of estroprogestinic drugs.

The association between VUR and APN is well known in children, while it is not clear in adults; therefore, there are no clear-cut indications to search for VUR in the context of APN.

In a study on 86 women affected by APN, only two cases with VUR were found [41]. In another study, 48 out of 603 women with APN (8%), who had recurrent episodes of APN, underwent a retrograde cystography [42]. Twenty-one had VUR: 12 of them (i.e., those who presented scars or ureteral duplication) received an anti-reflux endoscopy correction, after which no further recurrences developed in 11 women. Hence, these authors suggest using an anti-reflux procedure to treat patients with urinary tract abnormalities, such as double ureters, cortical scars, or abnormality of ureteral orifices, which may indicate VUR even in the absence of obvious signs of ongoing VUR. The hypothesis at the basis of this approach is the high probability of a preexisting or a transient VUR which can no longer be detected.

We usually perform retrograde cystography on patients with recurrent APN or those who present dilation of the urinary tract [11]. On the basis of these criteria, 43 patients in our series of 223 underwent retrograde urethrocystography: VUR was found in 9 of them (20.9%) [11].

Home treatment can be undertaken if the patient’s conditions are good, and fever and leukocytosis are mild.

Hospitalization is mandatory if the patient is suffering, vomits, presents signs of sepsis or risk factors, or a complicated form of APN.

Parenteral treatment may be started in hospital and continued at home.

Guidelines on APN treatment have been published by the IDSA [12], by the European Association of Urology [43], and by the Scottish Intercollegiate Guidelines Network [44].

In case of uncomplicated APN, the Guidelines suggest oral treatment with antibiotics at home or a rapid switch to oral treatment.

Non-severely ill patients may be treated by oral ciprofloxacin for 7 days, levofloxacin 750 mg per day for 5 days, or by trimethoprim/sulphamethoxazole (TMP/SMX) 320/1600 mg/day for 14 days if the sensitivity of the microorganism to this agent is known.

A single intravenous dose (ceftriaxone 2 g, gentamycin 3–5 mg/kg or a fluoroquinolone—i.e., ciprofloxacin 400 mg) may be administered before oral therapy.

More severely ill patients must be treated intravenously with a fluoroquinolone or an aminoglycoside with or without ampicillin, or a third-generation cephalosporin with or without an aminoglycoside or a carbapenem.

As regards gram-positive cocci, treatment with ampicillin/sulbactam (or amoxicillin/clavulanic acid) with or without an aminoglycoside is recommended.

Aminoglycosides have a wide spectrum of bactericidal action which is synergic with beta-lactam antibiotics. They are indicated in sepsis, in cases of suspected resistant gram-negative bacteria, and in association with beta-lactam antibiotics or fluoroquinolones until the microbes have been identified or in case of allergy to other antibiotics [45].

When an improvement is attained, antibiotic therapy may be administered orally. A fluoroquinolone or TMP/SMX is recommended (for gram-positive bacteria amoxicillin or amoxicillin/clavulanic acid).

Fever usually disappears within 72 h of beginning treatment. In a study performed in uncomplicated APN, fever disappeared in 26% of patients after 48 h and in 13% after 72 h [46]. Hence, the persistence of fever beyond 72 h does not necessarily require a change in therapy [47].