Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors and a predictive model

EJINME-03419; No of Pages 6 European Journal of Internal Medicine xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim

Original Article

Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors and a predictive model Joaquin Campos-Franco a,⁎, Cristina Macia a, Emilio Huelga b, Carla Diaz-Louzao c, Francisco Gude c, Rosario Alende a, Arturo Gonzalez-Quintela a a b c

Department of Internal Medicine, Hospital Clinico Universitario, Santiago de Compostela, Spain Department of Radiology, Hospital Clinico Universitario, Santiago de Compostela, Spain Department of Clinical Epidemiology Unit, Hospital Clinico Universitario, Santiago de Compostela, Spain

a r t i c l e

i n f o

Article history: Received 30 September 2016 Received in revised form 20 November 2016 Accepted 6 December 2016 Available online xxxx Keywords: Pyelonephritis Acute focal bacterial nephritis Acute lobar nephronia Adult Nomograms

a b s t r a c t Background: Acute focal bacterial nephritis (AFBN) is a complicated form of acute pyelonephritis (APN) characterized by single or multiple areas of localised infection in the kidney without liquefaction or abscess. Studies investigating AFBN in adults are scarce. Aim: The present study was aimed at evaluating the prevalence, associated factors, and presence of atypical clinical and radiological manifestations in adult AFBN patients. Also, we developed a clinical prediction model to evaluate the probability of AFBN in patients with APN. Methods: The clinical records of 377 patients (mean age 54 years, 74.0% females) admitted to a hospital over a 5year period with APN were reviewed. Results: A total of 57 cases of AFBN were radiologically identified (prevalence, 15.1%). Patients with AFBN were younger and displayed atypical manifestations more frequently than patients without AFBN; these included both clinical and radiological (pleural effusion, gallbladder wall thickening, fluid around the gallbladder, perirenal fluid, and ascites) manifestations. Patients with AFBN showed lower systolic blood pressure and needed more days of therapy to become afebrile, longer total duration of antibiotic therapy, and longer hospital stay than patients without AFBN. Contraceptive use was more frequent in patients with AFBN. A model based on five clinical variables showed good discrimination performance for the diagnosis of AFBN (Area under the curve, 0.77 (95% CI, 0.69–0.89)). Conclusions: Patients with AFBN frequently present with atypical clinical and radiological manifestations. Clinical presentation by means of a predictive model may predict the presence of AFBN. Patients with AFBN need more intensive therapy, which is followed by a favourable outcome. © 2016 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

1. Introduction Acute pyelonephritis (APN) is an infection of the upper urinary tract, including the renal pelvis, calices and parenchyma [1]. APN is common worldwide, accounting for many hospitalizations each year; among the female population, annual rate of inpatient pyelonephritis is estimated in 3–4 cases per 10.000 population [2]. Acute focal bacterial nephritis (AFBN) is a radiological diagnosis that was described by Rosenfield et al. in 1979 [3]. AFBN is a complicated form of APN characterized by single or multiple areas of localised infection in the kidney without liquefaction or abscess [3–6]. This condition is also known as acute lobar nephronia [3,6]. It has been proposed that AFBN is an intermediate disorder between APN and renal abscess [7,8]. Clinical ⁎ Corresponding author at: Department of Internal Medicine, Hospital Clinico Universitario, 15706 Santiago de Compostela, Spain. E-mail address: [email protected] (J. Campos-Franco).

presentation of AFBN is usually indistinguishable from the APN; therefore the diagnosis is often delayed and can lead to long-term morbidity (renal scarring, hypertension or renal dysfunction) [9]. Also, AFBN requires longer duration of antibiotic therapy [10]. Initially considered as rare, the widespread use of imaging studies allowed a better understanding and an ever-increased diagnosis of this condition. Most published studies collected cases in children [9– 19], with prevalence of AFBN ranging from 3.2% [18] to 8.6% in pediatric patients with febrile urinary tract infection (UTI) [9]. Another study excluding high risk children (those with recurrent UTI's, urogenital malformations, neoplastic conditions or immunodeficiency) found a 19.2% prevalence of AFBN [19]. Congenital malformations (particularly vesico-ureteral reflux) underlie in most children affected of AFBN [9, 17,18]. However, reports of AFBN in adult population are scarce [3,5, 20–33]. The largest series reported 30 cases over a three-year period [5]. The present study was aimed at evaluating the prevalence, associated risk factors, and presence of atypical clinical and radiological

http://dx.doi.org/10.1016/j.ejim.2016.12.002 0953-6205/© 2016 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002

2

J. Campos-Franco et al. / European Journal of Internal Medicine xxx (2016) xxx–xxx

manifestations in adult AFBN patients admitted to the hospital with APN. Also, we developed a clinical prediction model to evaluate the probability of AFBN and improve the clinical management of these patients. 2. Patients and methods 2.1. Study design and population This retrospective study included all patients who were hospitalized due to APN in the Internal Medicine Department of a university hospital between January 2011 and December 2015. The hospital covers area of approximately 500,000 inhabitants. Patients with APN are routinely admitted when there are signs of sepsis, renal failure, or complicated infection [1]. Using the International Classification of Disease, 9th revision (ICD-9), the computerized medical records of patients discharged from hospital with a diagnosis of APN, codes 590.10 (“acute pyelonephritis without lesion of renal medullary necrosis”), and 590.80 (“pyelonephritis, unspecified”) were reviewed. Patients with renal abscesses were excluded. A total of 401 episodes of APN requiring hospitalization occurred during the study period. Thirteen cases corresponded to repeated episodes in the same patients; in these cases only the first episode was considered. Eleven additional cases were excluded from the study because of lacking imaging studies (see below); these eleven patients were discharged uneventfully. Thus, the study included 377 patients with APN. Median age was 54 years (range, 15–96 years), with a 74.0% female proportion. 2.2. Main determinations Data regarding length of hospital stay, intensive care unit (ICU) admission, risk factors for infection, clinical presentation, imaging studies, microbiological, and laboratory results were obtained from the electronic clinical records of the patients. The following were considered as risk factors for UTI: diabetes mellitus, use of contraceptive pills, history of urolithiasis or prostatic disease, immunodeficiency, history of recurrent UTI, prior episodes of APN, permanent bladder catheter, recent urinary procedures (including prostate biopsy, cystoscopy, or urinary catheter placement performed in the last week, prior to the admission), and urogenital malformation. The use of antibiotics prior the admission was also recorded. Systolic arterial pressure, axillary temperature and heart rate at admission were recorded as well as the main clinical manifestations referred by the patients (flank or costovertebral angle pain, urinary symptoms, and nausea or vomiting). The lowest systolic blood pressure and highest heart rate registered during hospitalization were also registered. Periumbilical pain, right or left lower quadrant pain, and presence of Murphy's sign were also recorded as atypical manifestations of APN. Pyuria was defined as N10 white blood cells (WBC) per microscopic high power field. Urine and blood cultures were reviewed and positive cultures were defined using standard criteria. The determination of antimicrobial susceptibility and the presence of extendedspectrum beta-lactamases (ESBL) were performed using the guidelines of the Clinical and Laboratory Standards Institute (CLSI-Ref) [34]. Routine laboratory determinations (WBC, blood hemoglobin, platelet count, erythrocyte sedimentation rate (ESR), serum creatinine, lactate, fibrinogen, C-reactive protein (CRP), aspartate-aminotransferase (AST), alanine-aminotransferase (ALT), gamma-glutamyl transpeptidase (GGT), bilirubin, and albumin levels) were recorded. Serum samples were obtained during the first 72 h after admission. Duration of fever was considered as the total number of days with fever after the start of therapy. Reports and images of abdominal ultrasonography (US) and computed tomography (CT) were reviewed by a same expert radiologist. In our institution, US is performed routinely in patients with APN. If not contraindicated, contrast-enhanced CT scan is performed when (a) abdominal US reveals significant abnormality (obstructive uropathy,

lithiasis, or renal mass), (b) fever persists after 72 h of the starting of antibiotic therapy, or (c) clinical findings make necessary to rule out different diagnoses (mainly appendicitis or diverticulitis). Contraindications for CT and use of intravenous contrast agents included: a history of allergy to contrast agents, pregnancy and chronic or acutely worsening renal disease. Patients were classified as having AFBN (a) when US revealed renal focal mass(es) of decreased, or less frequently increased, echogenicity and decreased vascularity on doppler that disrupted corticomedullary definition; or (b) when CT scan revealed one or multiple wedge-shaped areas of decreased kidney density (striated nephrogram) after the injection of contrast medium. The remaining cases were classified as having uncomplicated APN (from now on, non-AFBN). Extrarenal manifestations of APN (including pleural effusion, ascites, or thickening of the gallbladder wall) [35,36] were registered. 2.3. Statistical analyses The Chi-square test was used for comparison of categorical variables and the t-Student test was used for comparison of continuous variables between groups. A p-value of b0.05 was considered statistically significant. Logistic regression analysis was used to estimate the probability of AFBN. Beginning with a model containing all potential covariates (clinical characteristics and risk factors included in Tables 1 and 2, respectively); Akaike Information Criterion (AIC) was used to select the best subset of predictor variables. Odds ratios (ORs) and their 95% confidence intervals (95% CI) were derived from the coefficients. Diagnostic performance of these models was assessed by constructing receiver operating characteristic (ROC) curves and was evaluated by calculating the area under the ROC curve (AUC). A nomogram with a visual scale method and prognostic information was also constructed. The point scale Table 1 Demographic and clinical characteristics of patients with acute focal bacterial nephritis (AFBN) and without AFBN (non-AFBN).

Age (years) Gender (female) Prior antibiotic therapy Lower urinary tract symptoms Flank pain Nausea and/or vomiting Atypical manifestationsa Temperature (°C) Systolic BP (mm Hg) Lowest systolic BP (mm Hg)a Heart rate (bpm) Pyuria (≥10 WBC/mL) Urine nitrite positivity Blood leukocytes (WBC × 103/μL) Blood hemoglobin (g/dL) Blood Platelets (×103/μL) ESR (mm/h) Plasma fibrinogen (mg/dL) Serum creatinine (mg/dL) Serum albumin (g/dL) Positive cultures Bacteremiab ESBL-producing microorganisms Duration of therapy (days) Hospital stay (days) ICU admission Duration of fever (days)

Non-AFBN (n = 320)

AFBN (n = 57)

p-Value

55 ± 22.2 236 (73.8%) 96 (30.0%) 211 (65.9%) 227 (70.9%) 124 (38.8%) 18 (5.6%) 37.9 ± 1.09 119 ± 21.6 103 ± 14.7 92 ± 17.2 307 (95.9%) 182 (56.9%) 14294 ± 6493 12.9 ± 1.7 234 ± 87.4 58 ± 27.7 550 ± 135 1.17 ± 0.67 3.4 ± 0.5 196 (61.3%) 72 (35.1%) 12 (3.8%) 13.4 ± 4.1 7.4 ± 5.3 14 (4.4%) 0.76 ± 0.9

44 ± 22.2 43 (75.4%) 15 (26.3%) 35 (61.4%) 34 (59.6%) 29 (50.9%) 14 (24.6%) 38.1 ± 1.03 110 ± 21.9 96 ± 13.3 95 ± 18.8 55 (96.5%) 24 (42.1%) 16501 ± 6.920 13.0 ± 1.4 225 ± 77.7 65 ± 23.5 522 ± 112 1.03 ± 0.44 3.3 ± 0.6 28 (49.1%) 9 (16%) 3 (5.3%) 18.6 ± 8.0 9.9 ± 8.5 8 (14%) 1.54 ± 1.2

b0.001 0.789 0.574 0.508 0.089 0.086 b0.001 0.185 0.006 0.004 0.324 0.512 0.046 0.020 0.661 0.467 0.058 0.165 0.151 0.570 0.086 0.115 0.483 b0.001 0.004 0.010 b0.010

Figures are means and standard deviation or as absolute number of patients and percentage (within parentheses). ICU: intensive care unit; ESBL: Extended-Spectrum Beta-Lactamases. ESR: erythrocyte sedimentation rate. BP: blood pressure. a Lowest systolic pressure (mm Hg) registered during hospitalization. b Blood cultures were performed in 277 patients of the non-AFBN group (86.5%) and in 56 patients of the AFBN group (98.2%).

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002

J. Campos-Franco et al. / European Journal of Internal Medicine xxx (2016) xxx–xxx Table 2 Distribution of risk factors between patients with acute focal bacterial nephritis (AFBN) and without AFBN (non-AFBN).

Diabetes mellitus Repeated episodes of UTI Previous APN episode Urolithiasis Urinary malformation Prostatic disease Permanent bladder catheter Recent urinary manipulation Immunodepression Oral contraceptive use Number of risk factors None One Two Three Four or more

Non-AFBN (n = 320)

AFBN (n = 57)

p-Value

61 (19.1%) 54 (16.9%) 27 (8.4%) 46 (14.4%) 32 (10%) 24 (7.5%) 11 (3.4%) 8 (2.5%) 37 (11.6%) 26 (8.1%)

9 (15.8%) 13 (22.8%) 5 (8.8%) 3 (5.3%) 1 (1.8%) 4 (7.0%) 0 (0.0%) 2 (3.5%) 2 (3.5%) 11 (19.3%)

0.558 0.280 1.000 0.059 0.042 0.898 0.383 0.652 0.066 0.009 0.228

101 (31.6%) 138 (43.1%) 56 (17.5%) 17 (5.3%) 8 (2.5%)

21 (36.8%) 23 (40.4%) 12 (21.1%) 1 (1.8%) 0 (0.0%)

Figures are absolute number of patients and percentage (within parentheses). APN: acute pyelonephritis. UTI: Urinary tract infection.

was assigned points to these variables in the nomogram based on the logistic regression model. All statistical analyses were carried out in R using the packages RMS (for fitting logistic models and building the nomogram), MASS (step AIC procedure) and pROC (for AUC and ROC curves). These packages are freely available at http://cran.r-project.org. 2.4. Ethical issues The study was performed in accordance with the ethical guidelines of the institution and was reviewed and approved (code 2016/473) by the Clinical Research Ethics Committee from Galicia (Spain). It conformed to the Helsinki Declaration and the Spanish regulations of confidentiality. 3. Results

b0.001

3.3. Imaging studies

The age of patients with AFBN (mean 44 years, range 18 to 86 years) was younger than that of patients without it (mean 55 years, range 15 to 96 years, p b 0.001). Female predominance was similar in patients with and without AFBN (Table 1). Use of antibiotic therapy prior to admission was similar in both groups (cefixime was the most frequently used

Table 3 Imaging studies and extrarenal findings in patients with acute focal bacterial nephritis (AFBN) and without AFBN (non-AFBN).

Studies performed US CT US and CT Pleural effusion Peri-renal fluid Fluid around the gallbladder Thickening of the gallbladder wall Ascites

255 (79.7%) 8 (2.5%) 57 (17.8%) 39 (12.2%) 28 (8.8%) 5 (1.6%) 3 (0.9%) 30 (9.4%)

23 (40.4%) 3 (5.3%) 31 (54.4%) 15 (26.3%) 19 (33.3%) 4 (7%) 3 (5.3%) 21 (36.8%)

3.2. Laboratory and microbiological data

p-Value

3.1. Clinical characteristics and risk factors

AFBN (n = 57)

antibiotic, data not shown). Interestingly, in patients with prior antibiotic therapy and positive cultures (n = 51), the antibiotic demonstrated to be well tailored in 29 cases. The distribution of risk factors in patients with and without AFBN is represented in Table 2. Contraceptive use was more common in patients with AFBN than in patients without it. Conversely, urolithiasis, urinary malformations, permanent bladder catheter, and immunodeficiency tended to be more frequent in the nonAFBN group (Table 2). Urethral stenosis (present in 9 patients without AFBN) was the most common malformation. Rheumatologic diseases (including rheumatoid arthritis and vasculitis, present in 16 patients without AFBN and one patient with AFBN) were the most common causes on immunodeficiency; cancer was present in 8 patients without AFBN; HIV infection was present in two patients without and in one patient with AFBN. The distribution of previous episodes of UTI, diabetes mellitus, prostate disease, and recent urinary manipulation was similar in patients with and without AFBN (Table 2). No hospitalizations of pregnant women occurred during the study period. Lower urinary tract symptoms and costo-vertebral angle pain were similarly reported by patients with and without AFBN. The frequency of chills was also similar in both groups (data not shown). Systolic blood pressure was significantly lower in patients with AFBN (Table 1). Atypical manifestations (periumbilical pain, right or left lower quadrant pain, or presence of Murphy's sign) were more prevalent in patients with AFBN than in patients without it (24.6% vs. 5.5%, p b 0.001; Table 1). Acute appendicitis or diverticulitis were suspected in 14 patients due to signs of peritoneal irritation. Acute cholecystitis was suspected in 14 other patients due to right upper quadrant pain and positive Murphy's sign. Of note, two of these patients (both young females) with AFBN underwent laparoscopic cholecystectomy. In both cases, the excised gallbladder showed no signs of cholecystitis. In both cases, US performed prior to surgery showed no signs of focal nephritis whereas a postsurgery CT revealed a multifocal pyelonephritis.

WBC and ESR were higher in patients with AFBN than in patients without it (Table 1). Blood hemoglobin, platelet count and plasma fibrinogen levels were similar in both groups (Table 1). Concentrations of blood lactate (only available in 181 patients), serum bilirubin, GGT, and alkaline phosphatase were similar in both groups (data not shown). One patient in the AFBN group suffered an ischemic hepatitis (maximum AST 7535 IU/L and ALT 3532 IU/L) in the course of septic shock and required prolonged ICU stay, haemofiltration and high-dose of noradrenaline. After exclusion of this case, no differences were observed in AST and ALT between patients with and without AFBN. Serum concentrations of CRP were only available in 81 patients (16 in the AFBN group and 65 in non-AFBN group) and tended to be higher in the AFBN group (p = 0.029, data not shown). Renal function (serum creatinine) was similarly preserved in patients with and without AFBN (Table 1). Positivity of urine cultures and bacteremia tended to be more frequent in non-AFBN group (Table 1). Urine nitrite positivity was also more frequent in patients without AFBN (Table 1). Escherichia coli was the most prevalent isolate in both groups, being present in 24 cases (78% of positive cultures) in the AFBN group and 171 cases (87% of positive cultures) in the non-AFBN group. The prevalence of extendedspectrum beta-lactamase producing bacteria was slightly higher in the AFBN group (Table 1).

Fifty-seven patients (15.1%) were classified as having AFBN. The remaining 320 patients with APN (84.9%) were classified as not having AFBN (non-AFBN). A comparison of demographics, clinical data, laboratory determinations, microbiological results, and imaging studies between patients with and without AFBN is summarized in Tables 1, 2, and 3.

Non-AFBN (n = 320)

3

0.005 b0.001 0.033 0.047 b0.001

The distribution of imaging studies was different in patients with and without AFBN (Table 3). A CT scan was more frequently performed in patients with AFBN, particularly after a US study. Extrarenal manifestations of APN such as pleural effusion, thickening of the gallbladder wall, presence of fluid around the gallbladder, perirenal fluid and presence of ascites were significantly more frequent in patients with AFBN

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002

4

J. Campos-Franco et al. / European Journal of Internal Medicine xxx (2016) xxx–xxx

than in patients without it (Table 3). When only patients who underwent a CT-scan were considered, the presence of ascites, perirenal fluid, and fluid around the gallbladder, were also significantly more frequent in patients with AFBN (data not shown). 3.4. Outcome Ceftriaxone was the most commonly used first-line antibiotic; it was administered to 41 patients (71.9%) with AFBN and 256 patients (80.0%) without AFBN. Piperacillin-tazobactam was the second preferred drug; it was used in 6 patients (5.6%) with AFBN and 20 patients (6.3%) without AFBN. Carbapenems were used in 4 patients with AFBN and 16 patients without it. Duration of therapy (summing up intravenous and oral therapy) was higher in patients with AFBN (Table 1). Duration of fever after initiating antibiotic therapy was longer in patients with AFBN than in patients without it (mean 1.54 days vs mean 0.76 days, respectively, p b 0.01). The length of hospital stay was longer in the AFBN group (Table 1). In the non-AFBN group, the duration of fever and the antibiotic therapy were longer in patients with bacteremia. The need for admission in the ICU was also more frequent in patients with AFBN than in patients without it (Table 1). Ureteral stent placement (double-J) was required in one patient from the non-AFBN group due to hydronephrosis. Only one patient died during hospitalization; this patient belonged to the non-AFBN group and died due to complications that were not related to APN (heart failure and respiratory acidosis). As a result of the imaging studies, four patients were incidentally diagnosed as having neoplastic conditions (renal cell carcinoma and colorectal cancer in two patients each). No renal abscesses were detected during follows up. 3.5. Predictive model for the diagnosis of AFBN A multivariate analysis of factors associated with AFBN among patients with APN is represented in Table 4. A history of contraceptive use, absence of urolithiasis, atypical manifestations (periumbilical pain, right or left lower quadrant pain, or presence of Murphy's sign), low systolic blood pressure, and duration of fever after initiating antibiotic therapy tended to be independently associated with the presence of AFBN. Using these data from logistic regression analyses, a nomogram was constructed in order to calculate the probability of having AFBN in patients with APN (Fig. 1). The diagnostic performance of the model for the diagnosis of AFBN is represented in Fig. 2. 4. Discussion In the present study, we found a 15.1% prevalence of AFBN in adult patients that were hospitalized with APN. Patients with AFBN were younger than patients without it. Patients with AFBN frequently displayed atypical clinical manifestations (periumbilical pain, right or left lower quadrant pain, and presence of Murphy's sign) and showed signs of more severe sepsis, as revealed for lower systolic blood pressure and higher need for intensive care unit admission than patients without AFBN. Patients with AFBN also displayed atypical radiological manifestations (pleural effusion, gallbladder wall thickening, presence of fluid around the gallbladder, perirenal fluid, and ascites) more frequently than patients without Table 4 Multivariate analysis (logistic regression) of factors associated with AFBN. Factor

Odds ratio (95% CI)

p-Value

Contraceptive use Urinary lithiasis Atypical manifestationsa Systolic blood pressure (mm Hg) Duration of fever (days)b

2.27 (0.93–5.23) 0.35 (0.08–1.06) 5.16 (2.21–11.96) 0.98 (0.96–1.00) 1.78 (1.37–2.33)

0.059 0.101 0.001 0.059 b0.001

All listed variables entered the equation. a Periumbilical pain, right or left lower quadrant pain, or presence of Murphy's sign. b After initiating antibiotic therapy.

AFBN. White blood cell counts were higher among patients with AFBN. Patients with AFBN needed more days of therapy to become afebrile, longer total duration of antibiotic therapy, and longer hospital stay than patients without AFBN. Anticonceptive use was more frequently reported by patients with AFBN than by patients without it. Conversely, traditional risk factors for UTI such as diabetes urolithiasis or urinary manipulation are not more common in patients with AFBN in our study. Clinical outcome under appropriate therapy is universally good in patients with AFBN. Probably, in agreement with other authors [12], AFBN is underdiagnosed and is not a rare entity in adults. Some factors have contributed to the limited number of cases reported in adults. Since AFBN is a radiological diagnosis, his prevalence relies on the use of imaging studies. US exploration is proposed as a routine in the evaluation of children with UTI [16,37], and this would explain the higher frequency of cases in children. However, adult population guidelines and recommendations from radiological societies recommend to reserve imaging studies for patients with atypical evolution or risk factors (diabetes, immunosuppression, equivocal diagnoses, history of lithiasis, persistent fever or no response to antibiotic therapy) [38]. Accordingly, patients without risk factors and with favourable evolution under antibiotic therapy can be discharged without imaging studies. Using these criteria, many cases of AFBN go unnoticed. In previous adult series and case reports, AFBN usually affects patients with risk factors. Diabetes is considered as the main predisposing factor in adults [5,39]. In the largest series of adults with AFBN, 87% of patients had diabetes or cirrhosis [5]. In the present study, diabetes and immunosuppression accounted for b20% of cases of AFBN; moreover, 36.8% of patients with AFBN had no risk factors. An additional problem for evaluating the frequency of AFBN in the literature is that some radiologists and physicians are reluctant to use the term AFBN, preferring complicated APN instead [9]. Cheng et al. proposed a distinction between simple AFBN and complicated AFBN based on clinical and CT findings [12,13]. Simple AFBN corresponds to a severe form of APN, but without evolution to renal abscess and usually responding to antibiotic therapy. Instead, complicated AFBN represents a more severe form that may be related with renal abscess. These patients had longer duration of fever after the start of antibiotic therapy compared with patients with simple AFBN [12]. One explanation is the presence of microabscesses or small areas of liquefaction [12]. This hypothesis is based in the observation of CT lesions compatible with complicated AFBN in the kidneys of patients with renal abscesses. The authors proposed that complicated AFBN could be a fulminant disease and can progress to renal abscess independently of the antibiotic therapy or the duration of fever [12]. One finding of our study is the association of atypical manifestations, either clinical (Murphy's sign) or radiological in patients with AFBN. Equivocal diagnoses have been reported in other cases series, including appendicitis and cholecystitis, leading to unnecessary procedures both in children and in adults [17,18,22,23]. Also, nephrectomy has been performed in cases presenting as renal mass to exclude cancer [3,23]. Two patients in our study underwent laparoscopic cholecystectomy; clinical examination revealed Murphy's sign and US demonstrated gallbladder thickening and fluid around it in both cases. Uncommon imaging CT findings have been reported in patients with severe APN (but it is unclear if any of these patients had features of AFBN) [35,36]. Perirenal fat infiltration, pelvic ascites, hepatic periportal tracking, thickened gallbladder wall and pleural effusion were reported, but some of them are nonspecific findings and had no relationship with clinical parameters [35,36,40,41]. Other studies proposed that periportal tracking and early inhomogeneous enhancement of the liver could be associated to severe forms of APN [41–43]. To the best of our knowledge, no studies have evaluated the prevalence of the extrarenal findings in patients with AFBN. In our study, such extrarenal manifestations as pleural effusion, thickening of the gallbladder wall or fluid around it, perirenal fluid and presence of ascites were associated with AFBN. However, the finding of extrarenal manifestations should be interpreted with caution.

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002

J. Campos-Franco et al. / European Journal of Internal Medicine xxx (2016) xxx–xxx

5

Fig. 1. A proposed nomogram for the diagnosis of acute focal bacterial nephritis (AFBN) in patients with acute pyelonephritis. Atypical: presence of atypical clinical signs (periumbilical pain, right or left lower quadrant pain, or presence of Murphy's sign). SBP: systolic blood pressure (in mm Hg). Duration: length of fever in days after initiating antibiotic therapy. Instructions: For three of the variables (lithiasis, contraceptives and presence of atypical clinical manifestations), the value “0” corresponds to “absent” and the value “1” is equal to “present”. Locate the factors associated with AFBN (e.g. lithiasis) on the respective axis and draw a line straight up to the point axis. Sum the points for each of the factors and locate the final sum on the total point axis. Draw a line straight down to find the patient's probability of having AFBN.

Some atypical radiological findings could therefore be due to accumulation of fluids in third spaces due to intravenous fluid overload in septic patients, further favoured by increased endotoxin-mediated vascular permeability and sepsis-related hypoalbuminemia [35]. Sepsis mediated hypoalbuminemia is caused by three main mechanisms: impaired hepatic production, persistent catabolism and increased transcapillary loss of albumin [44,45]. Cytokines and other inflammatory mediators are also involved in the increased transcapillary leakage to the interstitial compartment [46]. High levels of acute phase reactants (C-reactive protein levels, ESR or fibrinogen) and interferon-gamma have been reported in patients with AFBN [47,48]. Fluid overload and hepatic capsular distention could also explain the Murphy's sign; recently, the term transient secondary cholecystalgia has been proposed to describe a similar clinical picture in patients with congestive heart failure, positive sonographic Murphy's sign and suspected acute cholecystitis [49]. The study has limitations that should be acknowledged. It was conducted in a single center, thus restricting external validity. The retrospective design has obvious limitations for information. Information bias could be also induced by the higher frequency of CT scans performed among patients with AFBN. In addition, our study population was highly selected because it was based on hospitalized APN patients. Prevalence of AFBN could be overestimated because a significant number of APN patients are managed as outpatients. Another limitation is the absence of microbiological confirmation in many cases. It must be taken in account that 30% of patients were under antibiotic therapy at admission. On the other hand, strength of the study is the sample size and the comprehensive comparison of epidemiological, clinical, imaging, microbiological, and outcome characteristics of patients with APN with or without AFBN. In addition to atypical symptoms and imaging findings, patients with AFBN display higher WBC counts on admission and lower systolic blood pressure. Admission on ICU is more frequently indicated in patients with AFBN than in patients without it. Patients with AFBN need more days of antibiotic therapy to become afebrile, as

previously reported [10], and have longer hospital stay than patients without AFBN. Intriguingly, AFBN is associated with previous contraceptive use and tends to be inversely associated with traditional risk factors for UTI such as urolithiasis. By means of logistic regression analysis, we constructed a model to predict the probability of having AFBN among

Fig. 2. Receiver Operating Characteristics (ROC) curve and 95% confidence interval (grey area) of the prediction score for acute focal bacterial nephritis (AFBN) in patients with acute pyelonephritis.

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002

6

J. Campos-Franco et al. / European Journal of Internal Medicine xxx (2016) xxx–xxx

patients with APN. A history of contraceptive use, absence of urolithiasis, atypical symptoms, low systolic blood pressure, and duration of fever after antibiotic therapy may help to predict the presence of AFBN in patients with APN. This may serve to avoid unnecessary procedures, as well as to intensify therapy in a subgroup of patients that would need longer antibiotic therapy [10] and that frequently need admission in intensive care units. These observations, however, should be confirmed in future studies. Conflict of interest disclosure On behalf of all authors, the corresponding author states that there is no conflict of interest. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Ramakrishnan K, Scheid DC. Diagnosis and management of acute pyelonephritis in adults. Am Fam Physician 2005;71:933–42. [2] Czaja CA, Scholes D, Hooton TM, Stamm WE. Population-based epidemiologic analysis of acute pyelonephritis. Clin Infect Dis 2007;45(3):273–80. [3] Rosenfield AT, Glickman MG, Taylor KJ, Crade M, Hodson J. Acute focal bacterial nephritis (acute lobar nephronia). Radiology 1979;132(3):553–61. [4] Boam WD, Miser WF. Acute focal bacterial pyelonephritis. Am Fam Physician 1995; 52(3):919–24. [5] Huang JJ, Sung JM, Chen KW, Ruaan MK, Shu GH, Chuang YC. Acute bacterial nephritis: a clinicoradiologic correlation based on computed tomography. Am J Med 1992; 93(3):289–98. [6] Rathore MH, Barton LL, Luisiri A. Acute lobar nephronia: a review. Pediatrics 1991; 87(5):728–34. [7] Shimizu M, Katayama K, Kato E, Miyayama S, Sugata T, Ohta K. Evolution of acute focal bacterial nephritis into a renal abscess. Pediatr Nephrol 2005;20(1):93–5. [8] Wood GN, Kalra PA, Mamtora H, Waldek S, Donoghue DJA. Febrile patient with a renal space-occupying lesion evolving into renal abscess: acute focal bacterial nephritis. Nephrol Dial Transplant 1995;10(7):1248–50. [9] Cheng CH, Tsau YK, Chang CJ, Chang YC, Kuo CY, Tsai IJ, et al. Acute lobar nephronia is associated with a high incidence of renal scarring in childhood urinary tract infections. Pediatr Infect Dis J 2010;29(7):624–8. [10] Cheng CH, Tsau YK, Lin TY. Effective duration of antimicrobial therapy for the treatment of acute lobar nephronia. Pediatrics 2006;117(1):e84–9. [11] Cheng CH, Tsau YK, LH S, Lin CL, Lin TY. Comparison of urovirulence factors and genotypes for bacteria causing acute lobar nephronia and acute pyelonephritis. Pediatr Infect Dis J 2007;26(3):228–32. [12] Cheng CH, Tsau YK, Chen SY, Lin TY. Clinical courses of children with acute lobar nephronia correlated with computed tomographic patterns. Pediatr Infect Dis J 2009;28(4):300–3. [13] Cheng CH, Tsau YK, Lin TYI. Acute lobar nephronia the midpoint in the spectrum of upper urinary tract infections between acute pyelonephritis and renal abscess? J Pediatr 2010;156(1):82–6. [14] Bitsori M, Raissaki M, Maraki S, Galanakis E. Acute focal bacterial nephritis, pyonephrosis and renal abscess in children. Pediatr Nephrol 2015;30(11):1987–93. [15] Klar A, Hurvitz H, Berkun Y, Nadjari M, Blinder G, Israeli T, et al. Focal bacterial nephritis (lobar nephronia) in children. J Pediatr 1996;128(6):850–3. [16] Kline MW, Kaplan SL, Baker CJ. Acute focal bacterial nephritis: diverse clinical presentations in pediatric patients. Pediatr Infect Dis J 1988;7(5):346–9. [17] Seidel T, Kuwertz-Bröking E, Kaczmarek S, Kirschstein M, Frosch M, Bulla M, et al. Acute focal bacterial nephritis in 25 children. Pediatr Nephrol 2007;22(11):1897–901. [18] Uehling DT, Hahnfeld LE, Scanlan KA. Urinary tract abnormalities in children with acute focal bacterial nephritis. BJU Int 2000;85(7):885–8. [19] Yang CC, Shao PL, CY L, Tsau YK, Tsai IJ, Lee PI, et al. Comparison of acute lobar nephronia and uncomplicated urinary tract infection in children. J Microbiol Immunol Infect 2010;43:207–14. [20] Farmer KD, Gellett LR, Dubbins PA. The sonographic appearance of acute focal pyelonephritis 8 years experience. Clin Radiol 2002;57(6):483–7.

[21] Rigsby CM, Rosenfield AT, Glickman MG, Hodson J. Hemorrhagic focal bacterial nephritis: findings on gray-scale sonography and CT. AJR Am J Roentgenol 1986; 146(6):1173–7. [22] Nosher JL, Tamminen JL, Amorosa JK, Kallich M. Acute focal bacterial nephritis. Am J Kidney Dis 1988;11(1):36–42. [23] Li Y, Zhang Y. Diagnosis and treatment of acute focal bacterial nephritis. Chin Med J (Engl) 1996;109(2):168–72. [24] Montejo M, Santiago MJ, Aguirrebengoa K, Garcia B, Goicoetxea J, Martin A. Acute focal bacterial nephritis: report of four cases. Nephron 2002;92(1):213–5. [25] Iga R, Uchino H, Kanazawa K, Usui S, Miyagi M, Kumashiro N, et al. Acute focal bacterial nephritis in an occupational allergy. Lancet Infect Dis 2014;14(7):656. [26] Joss N, Baxter G, Young B, Buist L, Rodger RS. Lobar nephronia in a transplanted kidney. Clin Nephrol 2005;64(4):311–4. [27] Maeshiro T, Hokama A, Fujita J. Acute focal bacterial nephritis. BMJ Case Rep Jun 3 2014;2014. [28] Sawamura T, Nakada T, Sasagawa I, Yamaguchi T, Suzuki Y, Iijima Y. Imaging studies on acute focal bacterial nephritis. Int Urol Nephrol 1993;25(3):221–8. [29] Sieger N, Kyriazis I, Schaudinn A, Kallidonis P, Dietel A, Ganzer R, et al. Acute focal bacterial nephritis can lead to unnecessary invasive procedures: a report of three cases. Urol Int Dec 18 2015. http://dx.doi.org/10.1159/000441955 (PubMed PMID: 26681296). [30] Thomalla JV, Gleason P, Leapman SB, Filo RS. Acute lobar nephronia of renal transplant allograft. Urology 1993;41(3):283–6. [31] Yang CW, Kim YS, Yang KH, Chang YS, Yoon YS, Bang BK. Acute focal bacterial nephritis presented as acute renal failure and hepatic dysfunction in a renal transplant recipient. Am J Nephrol 1994;14(1):72–5. [32] Conley SP, Frumkin K. Acute lobar nephronia: a case report and literature review. J Emerg Med 2014;46(5):624–6. [33] Lee JK, McClennan BL, Melson GL, Stanley RJ. Acute focal bacterial nephritis: emphasis on gray scale sonography and computed tomography. AJR Am J Roentgenol 1980; 135(1):87–92. [34] Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; twenty-second informational supplement. Wayne, PA: CLSI; 2012. [35] Zissin R, Osadchy A, Gayer G, Kitay-Cohen Y. Extrarenal manifestations of severe acute pyelonephritis: CT findings in 21 cases. Emerg Radiol 2006;13(2):73–7. [36] Kim JS, Lee S, Lee KW, Kim JM, Kim YH, Kim ME. Relationship between uncommon computed tomography findings and clinical aspects in patients with acute pyelonephritis. Korean J Urol 2014;55(7):482–6. [37] Huang HP, Lai YC, Tsai IJ, Chen SY, Tsau YK. Renal ultrasonography should be done routinely in children with first urinary tract infections. Urology 2008;71(3):439–43. [38] Goel RH, Unnikrishnan R, Remer EM. Acute urinary tract disorders. Radiol Clin North Am 2015;53(6):1273–92. [39] Patterson JE, Andriole VT. Bacterial urinary tract infections in diabetes. Infect Dis Clin North Am 1997;11(3):735–50. [40] Talarico HP, Rubens D. Gallbladder wall thickening in acute pyelonephritis. J Clin Ultrasound 1990;18:653–7. [41] Zissin R, Kots E, Rachmani R, Hadari R, Shapiro-Feinberg M. Hepatic periportal tracking associated with severe acute pyelonephritis. Abdom Imaging 2000;25(3):251–4. [42] Vollmann R, Schaffler GJ, Spreizer C, Quehenberger F, Schoellnast H. Clinical significance of periportal tracking as an extrarenal manifestation of acute pyelonephritis. Abdom Imaging 2011;36(5):557–60. [43] Han GJ, Lee NK, Kim S, Kim TU, Song SH, Kim HS, et al. Septic liver: clinical relevance of early inhomogeneous enhancement of the liver in patients with acute pyelonephritis. Acta Radiol 2013;54(8):975–80. [44] Horowitz IN, Tai K. Hypoalbuminemia in critically ill children. Arch Pediatr Adolesc Med 2007;161(11):1048–52. [45] Nicholson JP, Wolmarans MR, Park GR. The role of albumin in critical illness. Br J Anaesth 2000;85(4):599–610. [46] Fleck A, Raines G, Hawker F, Trotter J, Wallace PI, Ledingham IM, et al. Increased vascular permeability: a major cause of hypoalbuminaemia in disease and injury. Lancet 1985;1(8432):781–4. [47] Kashiwagi Y, Kawashima H, Nara S, Ushio M, Nishimata S. Marked elevation of interferon-γ in acute focal bacterial nephritis. Clin Exp Nephrol 2012;16(4):656–7. [48] Yang WJ, Cho IR, Seong dH, Song YS, Lee DH, Song KH, et al. Clinical implication of serum C-reactive protein in patients with uncomplicated acute pyelonephritis as marker of prolonged hospitalization and recurrence. Urology 2009;73:19–22. [49] Desautels CN, Tierney DM, Rossi F, Rosborough TK. Case report: an unrecognized etiology of transient gallbladder pain in heart failure diagnosed with internist-performed point-of-care ultrasound. Crit Ultrasound J 2015;7(2).

Please cite this article as: Campos-Franco J, et al, Acute focal bacterial nephritis in a cohort of hospitalized adult patients with acute pyelonephritis. Assessment of risk factors a..., Eur J Intern Med (2016), http://dx.doi.org/10.1016/j.ejim.2016.12.002