a b s t r a c t

Purpose: Predicting medical outcomes for acute pyelonephritis (APN) in women is difficult. Delay in diagnosis and treatment often results in Rapid progression to circulatory collapse, multiple-organ failure, and death. The aim of this study was to investigate the value of procalcitonin level in women with APN at ED. Methods: We conducted a prospective study of women with APN presenting to the ED. The authors measured inflammatory biomarkers, and the severity of pyelonephritis was assessed by 4 Severity of disease classification system and stage of sepsis. We performed an analysis to assess the value of PCT for the prediction of 28-day mortality and disease severity.

Results: A total of 240 female patients with APN are included. Patients were divided into 4 groups on the basis of systemic inflammatory response syndrome criteria, organ dysfunction, and persistent hypotension. The median PCT level was higher in the septic shock group compared with other groups. Of the other inflammatory markers, only white blood cell count was significantly different among the groups, whereas high-sensitivity C-reactive protein level and erythrocyte sedimentation rate revealed no differences. The area under the curve for PCT in predicting 28-day mortality was 0.68. For predicting mortality, a cutoff value of 0.42 ng/mL had a sensitivity of 80% and a specificity of 50%. However, the disease Classification systems were demonstrated to be superior to PCT in predicting 28-day mortality.

Conclusions: Relative to other classic markers of inflammation, by distinguishing the severity of sepsis related to APN, PCT levels can provide additional aid to clinicians in disease severity classification and their decision of treatment at ED.

(C) 2013

Introduction

Acute pyelonephritis (APN) is one of the most common infections occurring in women. Greater than 50% of women contract APN at least once during their lifetimes. In the United States, more than 7 million women present as outpatients with symptomatic Urinary tract infections , more than 1 million visit emergency departments (EDs), and more than 100 000 are admitted to hospitals annually [1]. In South Korea, the annual incidence of APN is 36/100 000, with 10 of these patients requiring admission [2]. Although most patients with APN have minor symptoms and low mortalities, some patients progress rapidly to life-threatening states such as septic shock, multiorgan failure, and death.

? Prior presentations: Pan-Pacific Emergency Medicine Congress 2012, Seoul, Korea, October 2012.

?? Funding sources/disclosures: None.

* Corresponding author. Department of Emergency Medicine, College of Medicine, The Catholic University of Korea, Yeouido St Mary’s Hospital, 10, Yuksam-Ro, Yeongdeungpo-Gu, Seoul 150-713, Korea.

E-mail address: [email protected] (S.P. Choi).

Approximately 10% of the patients with APN visiting an ED require admission for treatment. Young, healthy women without complications or overall toxicity can be safely treated as outpatients, whereas admission is recommended for children, elderly individuals, and pregnant women because these groups have higher likelihoods of sepsis [3]. However, it is difficult to predict the progression to poor outcomes in patients without obvious immediate complica- tions. Although there have been reports of admission criteria based on the related risk factors, comorbidities, and severity of APN, low- risk patients who may be safely treated as outpatients are occasionally admitted for treatment [4]. Currently, clear regulations regarding the necessity for admission among patients presenting to an ED as well as methods of staging the severity of sepsis caused by APN are lacking.

Recent research has suggested that procalcitonin is a unique marker of inflammation in response to infection. In a clinical study of children, PCT was not only useful in the early diagnosis of APN but also was highly associated with Renal damage [5]. The measurement of PCT in patients with fever who present to ED is helpful for the diagnosis of the infection and for providing information regarding its prognosis

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2013.04.012

[6]. In addition, PCT values determined in the ED are useful to distinguish between mild-to-moderate sepsis and severe sepsis [7].

Prior research has mainly been performed with those patients who presented to an ED and who were suspected of having infections or severe diseases, and no studies on the role of PCT in differentiating mild-to-moderate sepsis, severe sepsis, and septic shock in a specific disease have been published. In particular, there are no previous investigations targeting female patients presenting with APN. The purpose of this study was to evaluate whether serum PCT values determined in the ED are useful in differentiating the stages of sepsis in patients diagnosed as having APN.

Methods

Study design

This investigation was a prospective observational study of female patients with APN presenting to an ED between May 2009 and April 2011. The study was conducted at a university hospital, which had a volume of approximately 30 000 patient visits per year. The study was approved by the institutional review board, with waiver of informed consent.

Study setting and population

Female patients who presented to an ED during the 2-year study period with urinary tract symptoms suggestive of APN were screened by an emergency medicine specialist for inclusion in the study. A medical record review was performed daily, along with a follow-up prospective observation of those patients who were agreed upon by 3 emergency medicine specialists. The final selection of participants in the last phase of the study was performed in consultation with a UTI specialist.

The eligibility criteria were as follows: (1) female patients, (2) age greater than 18 years, (3) body temperature greater than 38?C or the presence of fever and chills 24 hours before presentation to the ED, (4) acute onset with at least 1 sign and symptom of a UTI (dysuria, urgency, frequency, suprapubic pain, flank pain, or costovertebral- angle [CVA] tenderness on physical examination), and (5) a positive nitrite or Leukocyte esterase dipstick test result, subsequently confirmed by urinalysis with more than 10 [5] white blood cells (WBCs) per milliliter. Inclusion of patients in the study was confirmed with a diagnosis of APN based on a positive urine culture, with a single microorganism at 104 colony-forming units/mL or greater, from either a catheterized or midstream, clean-void Urine sample [8,9].

The exclusion criteria were as follows: (1) pregnant or breast- feeding patients; (2) a history of functional or structural urinary tract abnormalities; (3) recent treatment for urolithiasis or hydronephro- sis, the present use of indwelling catheters, or a history of nephrostomy; (4) patients on hemodialysis or Peritoneal dialysis;

(5) a history of kidney transplantation; (6) the presence of cancer, immunodeficiency, HIV, or other infections; and (7) a history of recent manual or instrumental urologic examination.

Baseline data collection and ED management

The treatment of patients who were diagnosed as having APN followed a recently published guideline [10]. The basic information regarding the patients was collected from the physician’s interview and the standardized medical record, which included demographic data (age and sex), coexisting illnesses, symptoms, physical examination findings, laboratory tests, radiographic findings, micro- biological analyses, and antibiotic use. The laboratory and radio- graphic investigations, with the exception of those required in the research protocol (ie, PCT levels, high-sensitivity C-reactive protein [hs-CRP] concentrations, WBC counts, and erythrocyte sedimentation

rates [ESRs]), were conducted at the discretion of each emergency medicine specialist.

The serum PCT levels, hs-CRP concentrations, WBC counts, and ESR were measured at the time of admission; venous blood was drawn from all patients and was transported immediately to the biochem- istry laboratory. The serum PCT analyses were performed by laboratory technicians in the Department of Laboratory Medicine using a chemiluminescence sandwich immunoassay (Roche Diagnos- tics, Zurich, Switzerland), with a direct measurement range of 0.02 to

100.00 ng/mL and a functional assay sensitivity of 0.06 ng/mL. The WBC counts were determined using a flow cytometer (TOA Medical Electronics, Kobe, Japan), according to the manufacturer’s instruc- tions. The ESR was determined using a quantitative capillary photometry method (Alifax, Padova, Italy). The hs-CRP level was measured using an automatic biochemical analyzer (Hitachi High- Technologies, Tokyo, Japan).

Statistical analyses

The patients’ demographic characteristics, coexisting illnesses, clinical signs/symptoms, and medical outcomes were analyzed. We performed ?² statistics or Fisher exact tests when appropriate for the qualitative variables and used the Kruskal-Wallis, analysis of variance, Wilcoxon rank sum test, and Student t test for the continuous variables with skewed distributions to compare the baseline patient characteristics.

We assessed the diagnostic performance of the assay using the receiver operating characteristic (ROC) curve, and we used the area under the ROC curve (AUC) to assess the overall Discriminatory power of the PCT, hs-CRP, ESR, and WBC count in predicting the severity of APN. For comparisons, we assessed the AUC of the Acute Physiological Assessment and Chronic Health Evaluation (APACHE) II, the New Simplified Acute Physiology Score (SAPS) II, the Sequential Organ Failure Assessment , score and the Mortality in Emergency Department Sepsis (MEDS) score. The AUC and its 95% confidence intervals (CIs) were estimated for each biomarker and for the severity of disease classification systems.

We estimated the optimal cutoff value that maximized sensitivity without a substantial loss of specificity. We also calculated the sensitivity, specificity, and the positive and negative predictive values (PPV and NPP) for each cutoff value for the PCT, hs-CRP, ESR, and WBC count data. We analyzed the Likelihood ratios as a measure of the extent to which the pretest odds were altered by the test results; a low LR (b 0.1) and a high LR (N 10) are considered useful in ruling out and ruling in a disease, respectively. P values less than .05 were considered to denote statistical significance. All analyses were performed using SAS version 9.1 software (SAS Institute Inc, Cary, NC).

Results

Characteristics of the study patients

A total of 389 women presented to the ED with fever or chills and exhibited at least 1 symptom or sign of UTI at the time of presentation during the study period; the patients were enrolled in the study based on their outcomes. Of these subjects, 105 met the exclusion criteria, and an additional 43 patients were excluded because of insufficient data. Of the remaining 250 patients, 10 could not be monitored for progress after 28 days, leading to the final number of 240 female patients with APN who were included in the analysis.

All patients in the study were women older than 18 years, and their basic characteristics are listed in Table 1. There were 54 patients with infections who did not meet the criteria for the diagnosis of Systemic Inflammatory Response Syndrome , 107 patients with sepsis, 53 patients with severe sepsis, and 26 patients with septic shock. The average age of the patients was 45.76 (+-15.86) years, and

Table 1

Baseline characteristics of 240 female patients presenting with APN

	All patients (n = 240)	Infection, no SIRS (n = 54)	Sepsis with SIRS (n = 107)	Severe sepsis (n = 53)	Septic shock (n = 26)
Age (y)a	45.76 +- 15.86	41.11 +- 14.27	46.52 +- 17.30	59.43 +- 12.42	63.23 +- 16.32
Comorbidities, n (%) Diabetes mellitusa	81 (33.75)	13 (24.07)	26 (24.30)	29 (54.72)	13 (50.00)
Hypertension	103 (42.92)	16 (29.63)	36 (33.64)	35 (66.04)	16 (61.54)
Urologic disease	18 (7.50)	4 (7.41)	5 (4.67)	5 (9.43)	4 (15.38)
Clinical sign Temperature (?C)a	38.08 +- 1.03	38.04 +- 0.82	38.14 +- 0.93	38.07 +- 0.91	38.12 +- 0.83
HR (beats/min)a	97.13 +- 20.69	85.67 +- 16.49	102.50 +- 16.24	96.58 +- 18.57	99.96 +- 35.56
MAP (mm Hg)a	89.25 +- 16.91	92.15 +- 15.46	91.91 +- 14.84	90.41 +- 16.54	69.88 +- 16.89
Clinical symptom, n (%) Dysuria	69 (28.25)	13 (24.07)	38 (35.51)	13 (24.53)	5 (19.23)
Frequency	62 (25.83)	12 (22.22)	36 (33.64)	11 (20.75)	3 (11.54)
Urgency	19 (7.92)	3 (5.56)	9 (8.41)	6 (11.32)	1 (3.85)
Suprapubic pain	19 (7.92)	2 (3.70)	11 (10.28)	4 (7.55)	2 (7.69)
CVA tendernessa	115 (47.92)	25 (46.30)	63 (58.88)	23 (43.40)	4 (15.38)
Laboratory value PCT (ng/mL)a	0.50 (0.13-4.06)	0.13 (0.07-0.47)	0.45 (0.13-2.22)	1.51 (0.26-11.50)	14.72 (0.65-70.60)
hs-CRP (mg/L)	88.03 (32.57-167.85)	39.70 (16.74-102.82)	84.98 (35.87-159.11)	96.18 (39.10-219.35)	129.43 (89.80-209.93)
ESR (mm/h)	54.00 (33.50-79.50)	46.00 (23.00-76.00)	55.00 (39.00-77.50)	56.50 (34.50-97.00)	55.00 (21.00-82.00)
WBC count (103/mm3)a	11.48 (8.49-14.68)	9.32 (6.98-11.14)	12.65 (10.02-14.95)	11.48 (8.37-15.92)	15.49 (7.30-25.95)
Urea nitrogen (mg/dL)a	17.00 (11.30-28.50)	13.15 (9.80-19.30)	14.00 (9.95-21.35)	27.20 (17.40-52.40)	44.25 (26.10-82.50)
Creatinine (mg/dL)a	0.99 (0.84-1.49)	0.90 (0.80-0.99)	0.95 (0.85-1.18)	1.60 (0.99-2.73)	2.02 (1.13-2.80)
e-GFR (mL/min/1.73m2)a	61.15 (37.30-77.74)	71.83 (59.76-83.41)	65.33 (50.04-79.21)	34.04 (19.24-62.31)	26.51 (17.55-48.95)
Previous antibiotics, n (%)	58 (24.17)	14 (25.93)	28 (26.17)	9 (16.98)	7 (26.92)

Data are shown as number (percentage) for dichotomous variables and median (IQR) or mean +- SD for continuous variables.

^a Value of statistical significance.

although there was no significant difference between the infection, the no-SIRS group, and the sepsis with SIRS group, the individuals in the severe sepsis and septic shock groups were older than those in the infection and sepsis groups. The most common medical history was hypertension (103 patients; 42.92%) followed by diabetes mellitus (81 patients; 33.75%) and cerebrovascular accident (31 patients; 12.92%), whereas 18 patients had previous Urologic diseases. Of the urinary tract symptoms and signs, flank pain or CVA tenderness on physical examination yielded significant differences among the groups. However, dysuria, urgency, frequency, and suprapubic pain did not exhibit significant differences among the groups. The median PCT value was 0.13 ng/mL (interquartile range [IQR], 0.07-0.49) in the infection group, 0.45 ng/mL (IQR, 0.13-2.31) in the sepsis group, 1.68 ng/mL (IQR, 0.26-11.50) in the severe sepsis group, and 14.72 ng/mL (IQR, 0.61-71.78) in the septic shock group; these results demon- strated an increase in PCT levels with an increasing severity of sepsis. Of the other inflammatory markers, only the WBC count was significantly different among the groups, whereas the hs-CRP level and ESR revealed no differences. The urea nitrogen, creatinine, and estimated glomerular filtration rate (e-GFR) were significantly different among the groups.

Table 2

Comparison between the mild and severe sepsis groups in APN

Mild sepsis group		Severe sepsis group	P
(n = 161)		(n = 79)
PCT (ng/mL)	0.28 (0.92-1.57)	3.61 (0.38-26.24)	b.05
hs-CRP (mg/L)	66.34 (28.56-146.62)	107.94 (38.94-213.73)	b.05
ESR (mm/h)	52.57 (33.25-76.83)	56.25 (33.25-93.50)	.27
WBC count (103/mm3)	11.27 (8.69-13.75)	12.05 (8.11-18.27)	.20
Urea nitrogen (mg/dL)	13.50 (9.88-20.45)	32.70 (18.70-57.53)	b.05
Creatinine (mg/dL)	0.93 (0.82-1.10)	1.72 (1.00-2.75)	b.05
e-GFR (mL/min per 1.73 m2)	67.72 (52.60-79.73)	31.55 (16.74-58.89)	b.05
MEDS	2.70 (0.50-5.46)	7.49 (5.55-11.04)	b.05
SAPS II	19.33 (12.18-25.04)	34.63 (27.06-42.50)	b.05
SOFA	0.73 (0.05-1.73)	4.52 (2.93-7.36)	b.05
APACHI II	6.33 (3.42-9.31)	16.67 (11.29-20.95)	b.05
Mortality, n (%)	6 (3.73)	14 (17.72)	b.05

Data are shown as median (IQR), unless otherwise indicated.

The comparison between the mild sepsis group (consisting of the infection, no SIRS, and sepsis with SIRS) and the severe sepsis group (consisting of the severe sepsis and septic shock) is analyzed. As predicted based on the basic characteristics of the patients, significant differences were observed between the 2 groups in the signs and symptoms of UTI. The levels of serum PCT demonstrated a statistically significant difference between the 2 groups. The median PCT value was 0.28 ng/mL (IQR, 0.92-1.57) in the mild sepsis group and 3.61 ng/ mL (IQR, 0.38-26.24) in the severe sepsis group. Unlike the previous results, the median hs-CRP values of the 2 groups were significantly different (66.34 mg/L [IQR, 28.56-146.62] and 107.94 mg/L [IQR, 38.94-213.73]). White blood cell count and ESR indicated no difference between the 2 groups. The markers of kidney function, including urea nitrogen, creatinine, and e-GFR, were all higher in the severe sepsis group (Table 2).

Medical outcomes and severity of disease classification system

A total of 204 of 240 patients were admitted for treatment, and 36 of these 204 received focused therapy in the intensive care unit (ICU; Table 3). The number of ICU admissions increased with a greater severity of sepsis, and 2 patients with severe sepsis and 1 patient with septic shock were transferred to another hospital for personal reasons. The median MEDS score of all patients was 4.67, and the scores increased from 2.34 to 11.64 with an increase in the severity of sepsis. However, there was no significant difference between the mild sepsis and the severe sepsis groups. The median SAPS II for all patients was 23.52, and the SAPS II scores for the infection, sepsis, severe sepsis, and septic shock groups were 18.86, 19.60, 32.00, and 41.25. The median SOFA score for all patients was 1.52, and the SOFA scores for the infection, sepsis, severe sepsis, and septic shock groups were 0.56, 0.84, 3.70, and 7.67. The median APACHE II score for all patients was 8.71, and the APACHE II scores for the infection, sepsis, severe sepsis, and septic shock groups were 5.29, 6.75, 15.00, and 19.33. Unlike the aforementioned MEDS score, the SAPS II, SOFA, and APACHE II classification systems demonstrated a statistically significant increase in the scores with an increase in the severity of sepsis. The overall mean length of stay was 4.6 (+-3.2) days and was longer for the septic shock group compared with the other groups.

Table 3

Medical outcomes and severity of disease classification system in APN

	All patients (n = 240)	Infection, no SIRS (n = 54)	Sepsis with SIRS (n = 107)	Severe sepsis (n = 53)	Septic shock (n = 26)
Admission, n (%)	204 (85.00)	36 (66.67)	92 (85.98)	51 (96.23)	25 (96.15)
ICU admission, n (%)	39 (16.25)	2 (3.70)	2 (1.87)	14 (26.42)	21 (80.77)
Length of stay (d)	4.6 +- 3.2	2.2 +- 2.1	4.3 +- 2.6	7.5 +- 3.8	9.1 +- 3.3
MEDS	4.67 (1.39-7.31)	2.34 (0.37-4.71)	2.91 (0.59-5.72)	6.24 (5.20-7.85)	11.64 (9.50-13.50)
SAPS II	23.52 (14.83-32.00)	18.86 (12.91-24.40)	19.60 (9.31-25.50)	32.00 (24.50-37.93)	41.25 (35.50-56.33)
SOFA	1.52 (0.36-3.64)	0.56 (0.00-1.35)	0.84 (0.11-1.90)	3.70 (2.48 -5.19)	7.67 (6.00-10.50)
APACHI II	8.71 (4.92-13.95)	5.29 (2.27-8.30)	6.75 (4.28-9.96)	15.00 (11.56-19.50)	19.33 (15.20-26.00)
Mortality, n (%)	20 (8.33)	2 (3.70)	4 (3.74)	3 (5.66)	11 (42.31)

Data are shown as number (percentage) for dichotomous variables and median (IQR) or mean +- SD for continuous variables.

Additional value of PCT for predicting the severity of APN

The inflammatory markers were analyzed for their ability to predict sepsis, severe sepsis, and septic shock in the patients with APN. As previously mentioned, the median PCT value yielded a statistically significant increase with an increase in the severity of sepsis. White blood cell count, urea nitrogen, creatinine, and e-GFR also demon- strated a statistically meaningful increase with increasing severity of sepsis. In contrast, hs-CRP levels were significantly different between the infection group and the severe sepsis and septic shock groups but were not significantly different among the sepsis, severe sepsis, and septic shock groups. Likewise, ESR was significantly different between the infection group and the sepsis and severe sepsis groups, but not among the sepsis, severe sepsis, and septic shock groups.

Compared with the other markers of inflammation used to predict sepsis, PCT showed the highest AUC (0.75 [95% CI, 0.69-0.80]), indicating a mild discriminative power (Fig. 1A). A cutoff value greater than 0.17 ng/mL had a sensitivity of 79% and a specificity of 65%. The AUC of the ROC curve of hs-CRP, ESR, and WBC count for sepsis was 0.67, 0.59, and 0.72, respectively. The AUC value of PCT for severe sepsis was 0.73 (95% CI, 0.67-0.78). Using serum PCT level of 3.60 ng/ mL as a cutoff, the sensitivity was 0.61 and the specificity was 0.84 (Fig. 1B). The AUC value of the PCT for septic shock was 0.76 (95% CI, 0.70-0.81). Using the PCT value of 4.33 ng/mL as a cutoff, the sensitivity was 0.75 and the specificity was 0.81 (Fig. 1C). The AUC value of hs-CRP, ESR, and WBC count to predict severe sepsis and septic shock was lower when compared with the AUC value of PCT.

There were 20 deaths during the study period, with 14 (17.72%) of 79 patients in the severe sepsis and septic shock groups, as well as 6 (3.73%) patients belonging to the other groups, who died within 28 days. The AUC values of the PCT, MEDS, APACHI II, SOFA, and SAPS II scores in predicting mortality were 0.68, 0.91, 0.88, 0.86, and 0.92, with the SAPS II yielding the highest AUC value. The disease classification systems demonstrated to be superior to PCT in predicting mortality (Fig. 2).

The multivariable Cox proportional hazard ratio was calculated. The adjusted relative mortality of 28 days for the severe sepsis group

compared with the sepsis group was 5.10 (95% CI, 1.96-13.28). This estimate increased to 12.55 (95% CI, 5.19-30.35) for the septic shock group compared with the sepsis group. Using the PCT value of 0.43 ng/ mL, hazard ratio was 3.78 (95% CI, 1.27-11.32) compared with a value that is lower than the cutoff.

A correlation analysis was performed to evaluate the relationship between the inflammatory biomarkers and the 4 scoring systems. The Pearson Correlation coefficients of PCT with the SOFA, SAPS II, APACHI II, and MEDS were 0.55, 0.46, 0.39, and 0.32, respectively (P b .05). The hs-CRP had the highest correlation coefficient with the SOFA score (0.34), and the WBC count yielded the highest correlation coefficient with the SAPS II score (0.39). The ESR did not demonstrate a meaningful correlation with any of the 4 scoring systems.

Predictive value of inflammatory biomarkers for bacteremia

Forty patients had positive blood cultures, 23 (14.29%) in the minor infection group and 17 (21.52%) in the major infection group. The most common organism found in the blood cultures was Escherichia coli (28 patients), followed by Klebsiella pneumonia and Proteus mirabilis. K pneumonia and P mirabilis were found only in the major infection group. The patients with positive blood cultures had a higher median PCT of 3.61 ng/mL (IQR, 0.34-14.47) compared with

0.30 ng/mL (IQR, 0.10-1.68) in those patients with negative blood cultures (Table 4). With a PCT cutoff level higher than 1.02 ng/mL, the sensitivity, specificity, PPV, and NPV were 0.67, 0.72, 0.50, and 0.84. The hs-CRP, WBC count, urea nitrogen, creatinine, and e-GFR were all higher in the bacteremia group. However, ESR revealed no difference between the 2 groups. To evaluate the capability of the inflammatory biomarkers in predicting the results of the blood cultures, the AUC was calculated. The calculated AUCs of PCT, hs-CRP, ESR, and WBC count were 0.72, 0.64, 0.56, and 0.60.

Discussion

Sepsis is a complicated, often misdiagnosed, heterogeneous condition that can lead to significant clinical outcomes. The severity

Fig. 1. The ROC curves of PCT, hs-CRP, ESR, and WBC count for the prediction of sepsis (A), severe sepsis (B), and septic shock (C) in women with APN.

Fig. 2. The ROC curve for the prognostic value of PCT and disease classification systems to predict mortality of APN.

of this condition is highlighted by its mortality rate, which is projected to be approximately 30% to 50% and is influenced by early detection and treatment. Consequently, the prompt confirmation and early treatment of patients with sepsis or who are at risk for sepsis are crucial. However, many patients may not exhibit a significant disease state at their Initial diagnosis.

Although APN is generally a condition with mild complications, it can progress rapidly to life-threatening conditions such as septic shock and multiorgan failure. Acute pyelonephritis has an overall mortality of approximately 0.3%; however, when accompanied by severe sepsis or septic shock, APN exhibits a dramatic increase in mortality and is closely linked to long-term admissions and multiple complications. Thus, physicians should carefully assess the severity of APN at its initial diagnosis in the ED. This study investigated whether the serum PCT level, hs-CRP concentration, ESR, and WBC count can distinguish the severity of sepsis in women older than 18 years who present with APN at ED. Procalcitonin was determined to be superior to the other markers of inflammation in identifying the severity of sepsis from APN. Furthermore, the ability of PCT to predict bacteremia among the patients with APN was superior to the other inflammatory markers, and this marker was most highly correlated with disease severity scoring systems such as APACHE II, MEDS, SAPS II, and SOFA. In contrast, PCT was not superior to these scoring systems in predicting mortality.

Procalcitonin has been suggested to be a useful marker for the early diagnosis of UTI and for the prediction of UTI complications, particularly in children [11]. The PCT level is more accurate than CRP level and WBC count for diagnostics, and it is useful for predicting renal scarring [12-14]. A meta-analysis has demonstrated that in young children with culture-positive UTI, a PCT value greater than 0.5 ng/mL not only is able to accurately predict the renal parenchymal involvement but also can distinguish APN from cystitis. Consequently, PCT can be used to stage the severity of disease in children with UTI, which can be useful in choosing appropriate therapies or additional investigations [15].

Relative to the number of studies on children with UTI, insufficient research has been performed on PCT in adult patients with APN. A

Table 4

Blood culture and laboratory values of APN

Positive (n = 40) Negative (n = 200) P

PCT (ng/mL) 3.61 (0.34-14.47) 0.30 (0.10-1.68) b.05

hs-CRP (mg/L) 121.60 (57.67-223.07) 69.89 (25.20-139.62) b.05

prior study comparing pyelonephritis and community-acquired pneumonia with a control group demonstrated a high sensitivity (but low specificity) of PCT [16]. Another study of PCT in adult female patients presenting with pyelonephritis at ED found no significant difference between those patients with Adverse medical outcomes within 28 days (0.51 ng/mL) and the patients without such outcomes (0.08 ng/mL) [17]. Other studies have determined that PCT, CRP, and midregional pro-atrial natriuretic peptide are not useful in selecting the site for treating adult patients with pyelonephritis [18].

The present study did not analyze the role of PCT in diagnosing APN; the median value of PCT was 0.50 ng/mL, which is similar to the value determined in a previous investigation (0.46 ng/mL). In contrast, the median hs-CRP level was 88.03 mg/L, which is lower than the previously reported value of 220 mg/L. Despite reports that CRP can supplement currently used clinical and laboratory parameters (used for guiding appropriate treatments) as a marker for prolonged admission and recurrence in patients with APN [19], this study found that hs-CRP, compared with the other inflammatory markers, was not useful in distinguishing the severity of sepsis. Although the definition of adverse outcomes mentioned in a study by Lemiale et al [17] differs from that of this study, the analysis from dividing the patients into the infection and sepsis groups and the severe sepsis and septic shock groups yielded PCT values that were lower than those values in the previous investigation of both groups. Furthermore, although PCT was determined to be less useful in predicting mortality of APN (AUC, 0.68), the power of PCT to predict severe sepsis and septic shock was superior to that of the other markers of inflammation. The PCT values of the deceased patients overlapped with the PCT values of the surviving patients, which were consistent with previously results [7,20]. Of the deceased patients in the mild sepsis group, 1 patient had a PCT value of 4.47 ng/mL, whereas the values for the other 5 patients were less than 1.0 ng/mL. The 3 deceased patients in the severe sepsis group all had PCT values less than 0.5 ng/mL. Five patients from the infection group exhibited PCT levels greater than 3.60 ng/mL, which was used as a cutoff value for severe sepsis. Three of these patients had bacteremia, and another had a history of diabetes mellitus.

A number of studies on the use of biomarkers in the early detection

of severe sepsis in ED patients have been conducted. A study comparing the use of PCT, CRP, and interleukin-6 levels found that PCT was more useful than CRP in detecting severe sepsis [21], and although PCT was not found to be superior to CRP in diagnosing bacterial infections, it was more useful than CRP in determining the severity of these infections [22]. Most studies on the link between PCT and sepsis have targeted clinically ill patients in general rather than patients with a specific disease. An analysis of PCT for the Early prediction of sepsis outcomes has revealed that the odds ratio for death at a PCT level greater than 0.12 ng/mL is 2.606 (P b .05). Of the 922 patients included in that study, 415 (45%) had APN [23]. However, contrasting results were obtained in this present study: that is, PCT levels were not useful in predicting the mortality of patients with APN. Furthermore, the AUC value was lower than the 4 other scoring systems that were used. At a cutoff value of greater than 0.43 ng/mL, the odds ratio, sensitivity, specificity, PPV, and NPV were 1.02 (95% CI, 1.00-1.03), 0.80, 0.50, 0.13, and 0.97, respectively.

There are a few limitations to this study. For example, the results were analyzed at a single Tertiary university hospital, and thus, the number of cases was limited. Similar to tertiary university hospitals, a relatively higher portion of patients with serious illnesses and/or advanced ages visit ED. This phenomenon resulted in a higher mortality rate in the present study compared with the mortality

rate of APN reported for the general patient population. Furthermore,

ESR (mm/h)

3 3

57.00 (40.00-80.00) 51.33 (28.17-79.25) .21

the use of antibiotics before presentation at the ED or before the onset

WBC count (10 /mm ) 12.88 (9.24-17.81) 11.15 (8.15-13.53) b.05

Urea nitrogen (mg/dL) 23.00 (15.13-39.40) 14.95 (9.63-24.10) b.05

Creatinine (mg/dL) 1.31 (0.97-1.84) 0.94 (0.82-1.19) b.05

e-GFR (mL/min per 1.73 m²) 42.98 (27.80-61.31) 67.54 (50.02-80.34) b.05

Data are shown as median (IQR).

of the symptoms was not considered. Although relatively few studies has been conducted in this area, the effect of antibiotic use on PCT levels could not be accounted for, and the timing of the peak levels of PCT (3 days after infection) was not reflected. Finally, this study

targeted patients presenting to an ED without considering patients visiting the outpatient clinic; therefore, the analysis was conducted based on the initial results of the ED presentation rather than on the peak levels of PCT after admission.

Conclusions

This study targeted female patients with APN who presented to an ED. Although previous research results suggest that the use of PCT as a prognostic factor for patients with pyelonephritis is limited, the results of this study suggest the usability of PCT from a different perspective. The PCT levels measured at the ED were not useful in predicting the mortality of the patients with APN. Nonetheless, relative to other classic markers of inflammation, by distinguishing the severity of sepsis related to APN, PCT levels can provide additional aid to clinicians in disease severity classification and their decision of treatment at ED. Further studies are needed to confirm these findings in a larger population.

Acknowledgments

This work was supported by the CMC Clinical Research Coordi- nating Center, The Catholic University of Korea. We wish to thank the center for assistance with the statistical analysis.

References

1. Foxman B, Klemstine KL, Brown PD. Acute pyelonephritis in US hospital in 1997: hospitalization and in-hospital mortality. Ann Epidermiol 2003;13(2):144-50.
2. Ki M, Park TS, Choi BY, et al. The epidermiology of acute pyelonephritis in South Korea, 1997-1999. Am J Epidermiol 2004;160(10):985-93.
3. Tenner SM, Yadven MW, Kimmel PL. Acute pyelonephritis: preventing compli- cations through Prompt diagnosis and proper therapy. Postgrad Med 1992;91(2): 261-8.
4. Elkharrat D, Chastang C, Boudiaf M, et al. Relevance in the emergency department of a decisional algorithm for outpatient care of women with acute pyelonephritis. Eur J Emerg Med 1999;6(1):15-20.
5. Pecile P, Miorin E, Romanello C, et al. Procalcitonin: a marker of severity of acute pyelonephritis among children. Pediatrics 2004;114(2):e249-54.
6. de Kruif MD, Limper M, Gerritsen H, et al. Additional value of procalcitonin for diagnosis of infection in patients with fever at the emergency department. Crit Care Med 2010;38(2):457-63.
7. Hausfater P, Garric S, Ayed SB, et al. Usefulness of procalcitonin as a marker of Systemic Infection in emergency department patients: a prospective study. Clin Infect Dis 2002;34(7):895-901.
8. Ramakrishnan K, Scheid DC. Diagnosis and management of acute pyelonephritis in adults. Am Fam Physician 2005;71(5):933-42.
9. Rubin RH, Shapiro ED, Andriole VT, et al. Evaluation of new anti-infective drugs for the treatment of urinary tract infection. Clin Infect Dis 1992;15(Suppl. 1): S216-27.
10. Gupta K, Hooton TM, Naber KG, et al. International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelonephritis in women: a 2010 update by the Infectious Diseases Society of America and the European Society for Microbiology and Infectious Diseases. Clin Infect Dis 2011;52(5): e103-20.
11. Ipek IO, Sezer RG, Senkal E, et al. Relationship between procalcitonin levels and presence of vesicoureteral reflux during first febrile urinary tract infection in children. Urology 2012;79(4):883-7.
12. Sheu JN, Chang HM, Chen SM, et al. The role of procalcitonin for acute pyelonephritis and subsequent renal scarring in infants and young children. J Urol 2011;186(5):2002-8.
13. Bressan S, Andreola B, Zucchetta P, et al. Procalcitonin as a predictor of renal scarring in infants and young children. Pediatr Nephrol 2009;24(6):1199-204.
14. Kotoula A, Gardikis S, Tsalkidis A, et al. Comparative efficacies of procalcitonin and conventional inflammatory markers for prediction of renal parenchymal inflam- mation in pediatric first urinary tract infection. Urology 2009;73(4):782-6.
15. Mantadakis E, Plessa E, Vouloumanou EK, et al. Serum procalcitonin for prediction of renal parenchymal involvement in children with urinary tract infections: a meta-analysis of prospective clinical studies. J Pediatr 2009;155(6):875-81.
16. Martinot M, Hansmann Y, De Martino S, et al. Procalcitonin in pyelonephritis and acute community-acquired pneumonia in adults. Presse Med 2001;30(22): 1091-6.
17. Lemiale V, Renaud B, Moutereau S, et al. A single procalcitonin level does not predict adverse outcomes of women with pyelonephritis. Eur Urol 2007;51(5): 1394-401.
18. Claessens YE, Schmidt J, Batard E, et al. Can C-reactive protein, procalcitonin and mid-regional pro-atrial natriuretic peptide measurements guide choice of in- patient or out-patient care in acute pyelonephritis? Biomarkers In Sepsis (BIS) multicentre study. Clin Microbiol Infect 2010;16(6):753-60.
19. Yang WJ, Cho IR, Seong do H, 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(1):19-22.
20. Suprin E, Camus C, Gacouin A, et al. Procalcitonin: a valuable indicator of infection in a medical ICU? Intensive Care Med 2000;26(9):1232-8.
21. Uusitalo-Seppala R, Koskinen P, Leino A, et al. Early detection of severe sepsis in the emergency room: diagnostic value of plasma C-reactive protein, procalcitonin, and interleukin-6. Scand J Infect Dis 2011;43(11-12):883-90.
22. Chan YL, Tseng CP, Tsay PK, et al. Procalcitonin as a marker of bacterial infection in the emergency department: an observational study. Crit Care 2004;8(1):R12-20.
23. Giamarellos-Bourboulis EJ, Tsangaris I, Kanni T, et al. Procalcitonin as an early indicator of outcome in sepsis: a prospective observational study. J Hosp Infect 2011;77(1):58-63.