Clinical Outcome and Prognostic Factors of Sepsis,Septic Shock and Prolonged Hospitalization,of Patients Presented with Acute Obstructive Pyelonephritis

Abstract

Objective:

To investigate prognostic factors of outcome of acute obstructive pyelonephritis (AOP).

Materials and Methods:

Patients with AOP were prospectively evaluated and logistic regression analysis was applied to identify factors associated with the duration of hospital stay and occurrence of sepsis and septic shock.

Results:

Based on CT scan findings, 62 patients were found to have AOP and subjected to emergency drainage. The main etiology of obstruction was lithiasis (70.9%). Double-J stent and percutaneous nephrostomy were introduced in 48 and 14 patients, respectively. Urosepsis and septic shock were diagnosed in 20 (32%) and 6 (9.7%) patients, respectively. None of the patients died of sepsis. In univariative analysis, older age, high neutrophils, increased serum creatinine, higher Charlson comorbidity index (CCI) score, any CCI score ≥1, diabetes mellitus (DM) longer operation time (OT), and multiresistant stains were risk factors of sepsis. Gender, type of drainage, laterality, white blood cell count, neutrophils rate >80%, C-reactive protein, and the presence of malignancy or lithiasis were not. Age, DM, and CCI score ≥1 were associated with prolonged hospitalization. None of the factors was associated with shock. In multivariative models, age (odds ratio [OR]: 1.09, 95% confidence interval [CI]: 1.02–1.16, p = 0.010), multiresistant strains (OR: 16.36, 95% CI: 1.97–135.71, p = 0.006), OT >20 minutes (OR: 1.03, 95% CI: 1.00–1.07, p = 0.048), and elevated creatinine (OR: 1.68, 95% CI: 1.001–2.84, p = 0.049) were independent prognostic factors of sepsis, and DM (OR: 30.8%, CI: 8.86%–52.8%, p = 0.007) was a prognostic factor of longer hospitalization.

Conclusions:

One-third of AOP patients will develop sepsis. Older age, elevated serum creatinine, longer OT presence of multiresistant strains, and DM are independent factors of worse outcome.

Introduction

One of the severe complications of the upper urinary tract obstruction is acute pyelonephritis, which is associated with high rates of hospital admissions, severe morbidity, or even death if not treated promptly. The incidence of the disease is anticipated to increase in the near future, considering that urolithiasis, which represents the main etiologic factor of acute obstructive pyelonephritis (AOP), has a higher prevalence nowadays than the past.^1

–4 However, not all of the patients with AOP will develop urosepsis or shock. Several authors outlined the stressing need to identify prognostic factors associated with poor outcome. In previous reports, older age, immunosuppression, decreased platelet count, impaired renal function, hydronephrosis, bacteriemia, and increased C-reactive protein (CRP) level determine the prognosis.^5,6 Moreover, the ideal method of renal decompression of an infected and obstructed pelvis is still a matter of debate.^7,8 In this study, we attempt to shed more light on the factors associated with the outcome and duration of hospitalization of patients found exclusively with AOP and no other associated pulmonary and intra-abdominal pathology analysis and to determine the best clinical practice for the management of the obstructed kidney.

Materials and Methods

Definitions

The clinical spectrum of acute pyelonephritis ranges from symptoms of cystitis to severe illness with sepsis and shock. It is often characterized by flank pain usually with chills, fever and unilateral or bilateral costovertebral angle tenderness. Dysuria, frequency, and urgency may coexist. Usually urine cultures are positive but ∼20% of patients host <10⁵ cfu/mL.⁹

Owing to the limitations in meeting all the aforementioned conditions, three basic criteria were used in the definition of pyelonephritis. Acute pyelonephritis was defined by at least two of the following: (1) auxilary temperature ≥38.3°C or chills, (2) pain or tenderness in the costovertebral angle, or pain at bimanual examination of the kidney, (3) pyuria or positive urine culture (≥10⁴ cfu/high-power field).^5,10,11

Sepsis is the life threatening dysfunction of organs and was defined as the systematic inflammatory response syndrome (SIRS) caused by infection. SIRS can be diagnosed any two of the followings: (1) temperature >38°C or <36°C, (2) heart rate >90 beats per minute, (3) respiratory rate >20 breaths per minute or partial pressure of carbon dioxide <32 mm Hg, and (4) white blood cells (WBCs) >12,000 cells per mm³ or <4000 cells per mm³ or ≥10% of immature forms.^10,12
–14

Septic shock is defined as sepsis with (1) signs of organ hypoperfusion: oliguria, lactic acidosis, impaired mental condition, or low platelet count (<10⁵/mm³) and (2) hypotension (systolic blood pressure <60 mm Hg or <80 mm Hg in patients with arterial hypertention) despite adequate fluid resuscitation with intravenously administered normal saline 0.9% at a rate >40 mL/kg or the need of vasoconstrictive agents (dopamine, epinephrine, norepinephrine) to maintain arterial pressure >60 mm Hg.^10

–13

Inclusion criteria

During a 3-year period (January 2016–December 2018), adult patients with pyelonephritis and suspected obstruction were immediately subjected to urinary tract imaging including abdominal CT scans. Those with confirmed AOP were recruited in the study and were prospectively followed up. Criteria for exclusion were diagnosis of nonobstructive pyelonephritis, recent urologic intervention, or previous treatment for AOP. Chronic pyelonephritis has distinct clinical and pathologic characteristics from those of acute pyelonephritis,⁹ and in the absence of symptoms and signs of acute infection, these patients were excluded from the study. Treatment of urinary tract infection (UTI) elsewhere or in outpatient basis, pregnancy, and transplantation were also criteria for exclusion. All the patients filled the relevant informed consent for their participation in the study. All of the methods of diagnostic evaluation and treatment and the study design were approved by the institute's ethics committee.

Evaluation and treatment

Clinical factors such as gender, age, etiology of obstruction, laterality, Charlson comorbidity index (CCI) score, diabetes mellitus (DM), method and operation time for renal decompression, and laboratory parameters such as number of WBCs, percentage of neutrophils, platelet count, CRP value, serum creatinine level, and detection of multiresistant bacterial strains were examined as potential prognostic factors for the primary and secondary end points.

For WBC count, two different upper limits (11 × and 15 × 10³/mm³) were examined. For CRP, the thresholds of 20 and 50 mg/dL were independently evaluated. The threshold for neutrophils rate was adjusted at 80%. Normal platelet count ranges from 150 to 400 × 10³/μL. CCI score was examined either as continuous or qualitative factor (0 vs ≥1). Upper normal creatinine level was set at 1.3 mg/dL of serum. Multiresistance was defined as strains resistant to three or more groups of broad spectrum antibiotics.¹⁵ Antibiotics were administered immediately after a free voiding urine sample had been delivered for urinalysis and culture. In severely ill patients or in those with low urine output, the sample was collected through a urethral catheter. As per our institute protocol, empiric antibacterial treatment consisted of carbapenems, or penicillins accompanied by aminoglucosides. If allergy to β-lactams was suspected, fluoroquinolons were administered instead. The treatment was initiated at the emergency department and changed accordingly to the sensitivity results.

The operation time (OT) was calculated from the administration of anesthesia till the removal of the endoscope or the rotation of the patient from the prone to supine position. Decompression of the affected renal unit was performed in no longer than 24 hours after presentation with either an endoscopically placed Double-J stent (DJS) or with a percutaneous nephrostomy (PN) placed with the patient in prone position. After the patient had been completely informed about the advantages and limitations of the two procedures, the decision regarding the method of decompression was reached by consensus between the physician and the patient, based on each one's preferences, the clinical condition, and the comorbidities of the patient. The duration and degree of obstruction were fundamental issues in the selection of treatment. All of the participants signed the relevant informed consent before the intervention.

Either procedure was performed under fluoroscopy. In PN, the initial needle puncture was aided by ultrasonographic guidance. Through the ureteral catheter or the nephrostomy needle, a second confirmative urine culture was collected directly from the affected kidney. All the interventions were performed under intravenously administered sedation or with epidural anesthesia. At the end of the procedure, a urethral catheter was placed in all of the patients.

Statistical analysis

The primary end point of the study was the occurrence of sepsis (diagnosed after the 2nd day of hospitalization) and secondary end points were the occurrence of septic shock and the duration of hospital stay. Shapiro–Wilk test was performed to test for normal distribution of continuous factors. The results of continuous factors are given as mean (± standard deviation), whereas not normally distributed factors are presented as median (± interquartile range [IQR]). Univariative logistic regression analysis was applied to identify factors related to sepsis and septic shock. Simple logistic regression was applied to identify factors related to the days of hospital stay. A logarithmic transformation was used for days of hospitalization before the application of logistic regression. Variables meeting the level of statistical significance were further analyzed as independent prognostic factors in multivariative logistic regression model. All tests were two tailed and statistical significance was established at 5% (p < 0.05). Data were analyzed using Stata™ (Version 13.0 MP; Stata Corporation, College Station, TX).

Results

Patients' characteristics are presented in Table 1. Overall, 62 patients enrolled in the study. The majority of them were females (58.1%) and were affected at a younger mean age compared with men. Fourteen patients (22.6%) suffered from cancer and 44 others (71.0%) from lithiasis.

Table 1.

Baseline Patients' Characteristics and Outcome

Characteristic/outcome	Mean (SD)/median [IQR]/count (%)
Gender
Females	36 (58.1%)
Males	26 (41.9%)
Age
Overall	64.5 (17.86)
Females	61.6 (17.80)
Males	68.4 (17.51)
Etiology of obstruction
Lithiasis	44 (70.9%)
Malignancy	14 (22.6%)
Other^a	4 (6.5%)
WBC count/mm³	13,550 [11,500–18,500]
Abnormal (WBCs <4000 or WBCs >10,000) WBC count patients	54 (87.1%)
Abnormal (WBCs <4000 or WBCs >15,000) WBC count patients	29 (46.8%)
% of neutrophils	83.6% (8.80%)
Platelet count ( × 10³/μL)	188 [140–305]
CRP (normal: <5 mg/L)	70.84 [36.08–158.77]
CRP >20	50 (80.6%)
CRP >50	39 (62.9%)
Serum creatinine level (mg/dL)	1.1 [0.8–1.9]
CCI (mean)	2 [0–5]
CCI (≥0) patients	44 (70.1%)
DM	23 (37.1%)
Multiresistant bacteria (pts)	11 (17.7%)
Renal decompression technique
DJS placement patients	48 (77.4%)
PN placement patients	14 (22.6%)
Hospital stay (days)	8 [6–11]
Sepsis events (pts)	20 (32.2%)
Septic shock events (pts)	6 (9.7%)
OT (minutes)	25 [15–35]

One patient with severe chronic prostate hyperplasia that caused ureterovisical junction obstruction, one with retroperitoneal fibrosis, and two with obstruction of ureteropelvic junction.

CCI = Charlson comorbidity index; CRP = C-reactive protein; DJS = Double-J stent; DM = diabetes mellitus; IQR = interquartile range; PN = percutaneous nephrostomy; OT = operation time; SD = standard deviation; WBC = white blood cell.

Renal decompression was performed with DJS and PN in 48 (77.4%) and 14 (22.6%) patients, respectively. The median duration of the operation, regardless of the method of decompression, was 25 minutes (IQR: 15–35). Only minor intra- and postoperative complications were reported. Four stented patients with prostate hyperplasia experienced prolonged hematuria and/or lower urinary tract symptoms (LUTSs). Regarding patients with PN, transient hematuria was recorded in two and lumbar pain in three others.

With respect to CRP levels, 80.6% and 62.9% of the patients exceeded the thresholds of 20 and 50 mg/dL, respectively. Two out of three cases (67.7%) had neutrophils ≥80%. CCI score ≥1 was present in 44 (70.1%) patients.

Urosepsis and septic shock were diagnosed in 20 (32%) and 6 (9.7%) patients, respectively. None of the patients died because of sepsis. Univariative analysis showed that older age (p = 0.004), increased neutrophils percentage, (p = 0.041), higher CCI score (p = 0.011), any CCI score ≥1 (p = 0.017), DM (p = 0.012), longer OT (p = 0.008), and multiresistant bacteria (p = 0.004) were correlated with sepsis. In contrast, gender, type of drainage, laterality, platelet count, neutrophils rate >80%, CRP level >20 or >50 mg/dL, and etiology of obstruction were not correlated with sepsis (Table 2). None of the examined factors was associated with septic shock (Table 3). Only age, DM, and CCI score ≥1 were related to days of hospitalization (Table 4).

Table 2.

Logistic Regression Models for Identification of Independent Factors Related to Sepsis

	OR	95% CI	p
Gender (RG = females)	2.2	0.74–6.50	0.154
Age	1.07	1.02–1.12	0.004
Intervention (RG = DJS)	2.69	0.79–9.17	0.113
Side (RG = right)	2.15	0.71–6.49	0.174
WBCs (RG = 4000 < WBCs <15,000)	2.21	0.74–6.54	0.153
WBCs (RG = 4000 < WBCs <11,000)	1.50	0.27–8.19	0.640
Neutrophils	1.08	1.003–1.16	0.041
Neutrophils (RG ≤80%)	3.85	0.98–15.2	0.054
CRP (for 10 U increase)	1.08	1.002–1.15	0.044
CRP (RG ≤20)	1.54	0.37–6.47	0.551
CRP (RG ≤50)	2.25	0.67–7.34	0.179
Ca (RG = no)	1.22	0.35–4.27	0.753
Lithiasis (RG = no)	0.93	0.29–3.00	0.908
CCI	1.33	1.07–1.67	0.011
CCI (RG = 0)	12.9	1.58–105.8	0.017
DM (RG = no)	4.22	1.37–13.07	0.012
Multiresistant	8.67	1.98–37.9	0.004
Creatinine	1.81	1.07–3.06	0.027
Platelet	0.99	0.99–1.003	0.531
OT	1.04	1.01–1.07	0.008

CI = confidence interval; OR = odds ratio; RG = reference group.

Table 3.

Logistic Regression Models for Identification of Independent Factors Related to Septic Shock

	OR	95% CI	p
Gender (RG = females)	2.22	0.34–14.32	0.403
Age	1.08	0.99–1.19	0.083
Intervention (RG = DJS)	2.50	0.37–16.7	0.344
Side (RG = right)	1.40	0.22–9.01	0.725
WBCs (RG = 4000 < WBCs <15,000)	1.79	0.28–11.53	0.541
WBCs (RG = 4000 < WBCs <10,000)	NA	NA	NA
Neutrophils	1.10	0.96–1.27	0.174
Neutrophils (RG ≤80%)	NA	NA	NA
CRP (for 10 U increase)	1.08	0.96–1.21	0.214
CRP (RG ≤20)	0.96	0.10–9.42	0.970
CRP (RG ≤50)	0.87	0.13–5.67	0.889
Ca (RG = No)	0.84	0.09–8.25	0.886
Lithiasis (RG = no)	1.70	0.18–16.3	0.646
CCI	1.24	0.89–1.72	0.198
CCI (RG = 0)	NA	NA	NA
DM (RG = no)	1.14	0.18–7.40	0.889
Multiresistant	1.17	0.12–11.67	0.890
Creatinine	1.17	0.61–2.26	0.635
Platelet	1.003	0.99–1.01	0.370
OT	1.02	0.99–1.06	0.169

NA = not available (all five patients with shock distributed to non-normal values group).

Table 4.

Simple Linear Regression Models for Identification of Independent Factors Related to Days of Hospitalization

	β coefficient^a	95% CI	p
Gender (RG = females)	−1.5%	−24% to 21%	0.892
Age	0.8%	0.16% to 1.4%	0.013
Intervention (RG = DJS)	22.6%	−3.4% to 48.6%	0.087
Side (RG = right)	0.09%	−22.5% to 22.7%	0.994
WBCs (RG = 4000 < WBCs <15,000)	11.1%	−11.0% to 33.2%	0.318
WBCs (RG = 4000 < WBCs <10,000)	−21.8%	−54.6% to 10.9%	0.188
Neutrophils	0.34%	−0.92% to 1.6%	0.586
Neutrophils (RG ≤80%)	8.1%	−15.6% to 31.9%	0.496
CRP (for 10 U increase)	0.73%	−0.70% to 2.2%	0.314
CRP (RG ≤20)	13.6%	−14.4% to 41.5%	0.336
CRP (RG ≤50)	16.4%	−6.2% to 39.1%	0.152
Ca (RG = no)	1.4%	−25.2% to 28.0%	0.918
Lithiasis (RG = no)	1.3%	−23.2% to 25.8%	0.917
CCI	4.1%	−0.11% to 8.2%	0.056
CCI (RG = 0)	25.5%	1.9% to 49.1%	0.035
DM (RG = no)	36.2%	15.1 to 57.3	0.001
Multiresistant	16.2%	−12.6% to 45.1%	0.265
Creatinine	8.1%	−1.0% to 17.3%	0.080
Platelet	−0.02%	−0.01% to 0.06%	0.542
OT	0.15%	−0.39% to 0.70%	0.573

% change of days of hospitalization for a unit increase of the independent factor.

In multivariative analysis, age (odds ratio [OR]: 1.09, p = 0.010), multiresistant strains (OR: 16.36, p = 0.006), OT >20 minutes (OR: 1.03, p = 0.048), and elevated creatinine (OR: 1.68, p = 0.049) were independent prognostic factors of sepsis (Table 5), and DM (OR: 30.8%, confidence interval [CI]: 8.86%–52.8%, p = 0.007) was an independent factor correlated with longer hospitalization (Table 6).

Table 5.

Multivariate Logistic Regression Model for Sepsis as Outcome

Factor	OR	95% CI	p
Age	1.09	1.02–1.16	0.010
Multiresistant (RG = no)	16.36	1.97–135.71	0.006
OT	1.03	1.00–1.07	0.048
Creatinine	1.68	1.001–2.84	0.049

Table 6.

Multivariate Logistic Regression Model for Logarithmic Transformation of Hospitalization

Factor	OR	95% CI	p
DM (RG = no)	30.8%	8.86%–52.8%	0.007

A comparative analysis between the two methods of decompression was performed with respect to the OT. It was revealed that the median OT was longer for the PN than for the DJS placement (35 minutes IQR [25–45] vs 20 minutes IQR [15–32.5], respectively, p = 0.008). Logistic regression analysis with sepsis as the outcome revealed that for the subgroup of DJS patients, it is 4% more likely to experience sepsis for 1 minute increase of OT (OR: 1.04, p = 0.015). In contrast for the PN subgroup, there is no statistically significant association between OT and sepsis (OR: 1.85, p = 0.363).

Discussion

AOP, despite the prompt and appropriate management, may progress to life-threatening urosepsis and septic shock with a mortality rate up to 9%. Should decompression of the obstructed renal unit be not performed at all, the risk of death increases by 2.6 times.¹⁶ Likewise, delayed interventions (performed later than 48 hours after the patient's admission) resulted in statistical significant increase in the mortality rate compared with that of early decompression (0.47% vs 0.16%, p = 0.02). Early intervention is considered an independent prognostic factor of decreased inpatient mortality (OR: 0.43, 95% CI: 0.19–0.98, p = 0.044).¹⁷

One of the authors' main concerns was the ideal time to evaluate patients for the occurrence of sepsis and septic shock. We estimated that the 2nd day of hospitalization was a reasonable time. It was a combination of the early evaluation that was necessary for patients with severe illness and at the same time it was an adequate time span as to allow the patient to overcome the surgical stress of renal drainage and the effect of anesthesia, both of which may influence and potentially bias our results. Patients critically ill were of course evaluated earlier (at the discretion of the responsible physician). However, classification of patients into sepsis, septic shock, or no sepsis group was based upon the findings at the 2nd day of hospitalization or later on.

In this study, despite the prompt drainage performed in <24 hours after admission, one-third of the patients eventually developed sepsis after 48 hours of hospital monitoring. The benefit of early drainage is evident in the rates of septic shock and death, which are as low as 9.7% and 0%, respectively. Compared with other series with similar study population, the percentage of shock ranges from 24 to 41.6%.^1,5,6,10 Timely decompression may increase the concentration of antibiotics within the kidney,¹¹ prevent or delay the inflammation process, and prohibit the development of severe sepsis and shock, leading to optimal results.¹⁸ Irrespective of the method of decompression, a postoperatively placed urethral catheter facilitated the drainage of the obstructed kidney. Even after a correctly placed and normally functioning PN, a urethral catheter was important for the first few days. Apart from accurate monitoring the urine output and early detection of hematuria, it also contributed to the maximum drainage of the affected renal unit, considering that a small but yet considerable urine volume may still be drained through the normal route.

The ideal method of drainage is still a matter of debate. An earlier study favors PN placement for the treatment of hydronephrosis,⁸ but recent findings seem to favor equally DJS and PN.¹ Compared with the latter, retrograde stent placement is less invasive, with negligible risk for hemorrhage and the patient does not need to carry a urine bag. This benefit is counterbalanced by the risk of increased LUTS and lower UTI.^7,19,20 Although the risk of major complications with PN, such as hemorrhage or perforation of adjacent organs, is overall low (0%–8%),⁷ it should be weighted in critically ill patients.²¹

Nevertheless PN has some advantages. Patients subjected to PN do not necessarily need to receive general anesthesia and be exposed to its side effects.⁸ This is important in severely ill patients with multiple comorbidities. PN may also be performed in the radiology suite (instead of the operation theater) by the interventional radiologist.²² This may prevent any delays or cancellations of the emergency operations list and may have a positive impact on the urologists and surgical staff because of the avoidance of additional stressful work.

Besides the aforementioned factors, in the present series, the method of choice for renal decompression was based primarily on the risk of failure of retrograde placement as determined by the medical history and the ongoing clinical condition. The type, duration, and degree of the obstruction also contributed to the determination of treatment choice. For example, in cases of long-lasting severe dilatation on the grounds of advanced malignancy of intrapelvic organs, the odds of failure of DJS favor the percutaneous renal drainage. In total, 91% (40/44) of the lithiasic patients but only 64% (8/14) of those with malignancy were effectively treated with DJS. In contrast, bleeding disorders or consumption of antiplatelet agents makes us reluctant to attempt a percutaneous intervention because of the risk of hemorrhage. Moreover, a DJS will also facilitate an upcoming endoscopic lithotripsy for definite treatment of urolithiasis. Every patient was informed about the clinical conditions, the available methods of treatment, the advantages and limitations of each technique, and the odds of effective placement before giving his or her consent.

In statistical analysis, neither the type of intervention nor the etiology of obstruction was an independent prognostic factor of the outcome. In our cohort of patients, DJS and percutaneous nephrostomy (PN) seem to be equally efficient with regard to sepsis (p = 0.113), septic shock (p = 0.344), and days of hospitalization (p = 0.087). However, based on the results of multivariative analysis, the duration of operation was correlated with the occurrence of sepsis but not with the septic shock or hospitalization. To the best of our knowledge, this is the first report on the role of OT in the outcome of AOP. Furthermore, in the subgroup analysis comparing the two decompression interventions, OT was only associated with the outcome of sepsis in the DJS subgroup and not in the PN group; this may reflect the smaller size of the PN group though.

The longer duration of decompression is attributed to difficulties in gaining access to the pelvis either with DJS or with PN. The attempt to manipulate and bypass impacted stones during DJS placement or the multiple attempts to access a poorly dilated collecting system during PN or DJS insertion are common causes for prolonged OT. Thus, high-volume instillations of contrast are likely to increase the intrarenal pressure and cause backflow of bacteria and toxins into the lymphovascular system,¹⁸ promoting the postoperative occurrence of sepsis. Although both treatments appear to be comparable for the occurrence of sepsis, a quick (<20 minutes) stenting procedure appears to be a more safe and effective management than a prolonged DJS placement.

Considering the mentioned, the right interventional choice, based on patients' clinical characteristics along with preference, is crucial for the outcome. If a DJS insertion is expected to take long, for example, in a patient with a really large impacted proximal ureteral stone with Grade 3 hydronephrosis, then an attempt for PN insertion seems safer and preferable.

It should be noted, however, that in some centers, the equipment and experience to quickly and safely perform PN are lacking. Elsewhere this procedure is performed by interventional radiologists instead of urologists, and they may not be readily available in the operation theater. In such circumstances, the only option to effectively treat AOP, should the placement of DJS fail, is to transfer the patient to another medical center where the placement of nephrostomy tube is feasible. This delay in offering an effective treatment increases the cost of health care and most importantly jeopardizes the patients' health condition. Owing to the small number of septic shock events, the role of OT and the type of intervention in the occurrence of shock cannot be evaluated.

In a previous report, Yoshimura et al. examined Karnofsky performance status (KPS) as an index of comorbidity in lithiasic obstruction. They concluded that comorbidities are independent risk factors of urosepsis, requiring immediate drainage.¹ However, the KPS as long as the Eastern Cooperative Oncologic Group tool has been evaluated and used primarily in oncologic populations.^23
–25 Considering that the majority of our patients (77.4%) had nonmalignant disease, these tools may not be ideal for our study population.

Therefore, we preferred the CCI score for preoperative evaluation of our patients. This tool offers a reliable estimation of the risk of postoperative complications and mortality related to the patient's comorbidities (i.e., diabetes, stoke, and myocardial infarction).^26,27

In univariative models, we demonstrated that higher CCI score increases the risk of sepsis by 33%. In addition, a patient with any score ≥1 is ∼13 times more likely to develop sepsis, which corresponds to the highest OR among the examined factors for sepsis (Table 2). Surprisingly, these results were not confirmed in multivariative logistic regression analysis. This inconsistency could be attributed to the moderate sample size. In contrast, a Taiwanese study failed to show any association of CCI score with severe sepsis in intensive care unit hospitalized patients. In this study, however, the group of patients was different from ours, considering that sepsis was associated mainly with pulmonary or abdominal diseases, whereas UTIs ranked fifth in the list of causative factors.²⁸

It is considered that patients with DM are vulnerable in developing severe pyelonephritis, sepsis, or shock.^5,10,11 Irrespective of others, our results did not confirm that DM is a statistically significant risk factor for the outcomes of sepsis or shock in multivariative logistic regression.

Not surprisingly, the presence of bacteria resistant to multiple antibiotics was an independent prognostic factor of urosepsis (Table 5). A potential explanation is that unless the sensitivity tests become available, the initial empiric antibiotic treatment delivered at the crucial early period of AOP is ineffective and the risk of sepsis and septic shock is high.¹⁸ Until the administration of targeted antibacterial therapy, early drainage of the affected renal unit is the only effective measure to prevent dissemination of bacteria and toxins and preserve as much of the renal function as possible.

Previous reports have shown that older age, female gender, poor performance status, hyperbilirubinemia, thrombocytopenia (platelet <100–120 10³/mm³), recent instrumentation, bacteriemia, and the grade of hydronephrosis (ectasia grade ≥II) were associated with sepsis and severe sepsis.^1,5 Poor performance status, positive blood culture, DM, thrombocytopenia, and low serum albumin level (platelet <100–120 10³/mm³) were correlated with septic shock.^5,6,10

We also provided robust evidence in multivariative models that older age and elevated serum creatinine also determine the prognosis of urosepsis (Table 5). Contrary to previous reports, in this study, the number of WBCs, the CRP level, and the platelet count did not. Moreover, none of the examined factors was correlated with shock, probably because of the limited number of relevant cases (9.6%).

With respect to the duration of hospital stay, older age (p = 0.013), DM (p = 0.001), and CCI score ≥1 (p = 0.035) were significant prognostic factors in univariative models (Table 4). Multivariative analysis, however, revealed that only DM was an independent risk factor for prolonged hospitalization (Table 6). Those with DM and perhaps older patients with coexisting chronic diseases may need more time to recovery after AOP despite early renal drainage, which explains our finding of longer stay in the hospital.

The main limitation of this study is the use of older definitions of sepsis and septic shock, instead of the recently introduced sepsis-related organ failure assessment (SOFA) and quick SOFA criteria.^14,29 During the recruitment period, the SOFA criteria were not popularized and reclassification of the patients after enrolment was not feasible. Furthermore, imaging of the urinary tract was performed in emergency settings, and when obstruction was confirmed, we proceeded to renal decompression without any time delay. Theoretically, a percentage of patients could have a favorable outcome without any drainage. The practice of early intervention (within 24 hours) may be unnecessary in these cases and perhaps represents an overtreatment in those with partial (incomplete) obstruction of recent onset.

Despite these limitations, our study revealed that earlier decompression (within the first 24 hours) improves the outcome by minimizing the rates of septic shock and death in AOP settings to 9.7% and 0%, respectively. We also revealed a strong correlation of the patients' age, multiresistant strains, OT, and creatinine with sepsis and of DM with the duration of hospital stay. To the authors' opinion, despite the lack of confirmation in multivariative settings, comorbidities (especially if they have no associated pulmonary or intraabdominal pathology analysis) could also be considered as risk factors because of their high degree of correlation with sepsis and hospital stay in univariative models. Our findings, however, warrant further evaluation by well-designed randomized controlled trials.

Considering that the main etiologic factors of obstruction were lithiasis and malignancies, an educational campaign should be undertaken by the national health care authorities aiming to decrease the incidences of these two diseases in the general population. Emphasis should also be given on the early treatment of urolithiasis as a measure to prevent its potentially catastrophic complications.

In the future, our findings, along with those from other similar publications, might be useful in the formation of a prognostic model of AOP. Classification of patients into high- and low-risk group for sepsis and septic shock might be helpful in defining those patients who will experience the maximum benefit from an early and aggressive interventional therapy.

Footnotes

Acknowledgment

Sincere thanks to Mr. George Tsampalas, MSc in biostatistics, for his valuable contribution in the conduct of statistical analyses.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

Abbreviations Used

References

Yoshimura

, Utsunomiya

, Ichioka

, Ueda

, Matsui

, Terai

. Emergency drainage for urosepsis associated with upper urinary tract calculi. J Urol, 2005; 173:458–462.

Trinchieri

Epidemiology of urolithiasis: An update. Clin Cases Miner Bone Metab, 2008; 5:101–106.

Liu

, Chen

, Liao

, et al. Epidemiology of urolithiasis in Asia. Asian J Urol, 2018; 5:205–214.

Sammon

, Ghani

, Karakiewicz

, et al. Temporal trends, practice patterns, and treatment outcomes for infected upper urinary tract stones in the United States. Eur Urol, 2013; 64:85–92.

Ruiz-Mesa

, Marquez-Gomez

, Sena

, et al. Factors associated with severe sepsis or septic shock in complicated pyelonephritis. Medicine, 2017; 96:e8371.

Kamei

, Nishimatsu

, Nakagawa

, et al. Risk factors for septic shock in acute obstructive pyelonephritis requiring emergency drainage of the upper urinary tract. Int Urol Nephrol, 2014; 46:493–497.

Pearle

, Pierce

, Miller

, et al. Optimal method of urgent decompression of the collecting system for obstruction and infection due to ureteral calculi. J Urol, 1998; 160:1260–1264.

Mokhmalji

, Braun

, Martinez Portillo

, Siegsmund

, Alken

, Köhrmann

. Percutaneous nephrostomy versus ureteral stents for diversion of hydronephrosis caused by stones: A prospective, randomized clinical trial. J Urol, 2001; 165:1088–1092.

Schaeffer

, Schaeffer

Infections of the urinary tract, Section V, Chapter 10. In: Wein

, Kavoussi

, Novic

, Partin

, Peters

, eds. Campbell-Walsh Urology, 10th Edition. Philadelphia, PA: Saunders, 2012, pp. 256–258, 295–297.

10.

Tambo

, Okegawa

, Shishido

, Higashihara

, Nutahara

. Predictors of septic shock in obstructive acute pyelonephritis. World J Urol, 2014; 32:803–811.

11.

Yamamoto

, Fujita

, Nakazawa

, et al. Clinical characteristics and risk factors for septic shock in patients receiving emergency drainage for acute pyelonephritis with upper urinary tract calculi. BMC Urol, 2012; 12:4.

12.

Levy

, Fink

, Marshall

, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med, 2003; 31:1250–1256.

13.

Wagenlehner

FME

, Pilatz

, Weidner

, Naber

. Urosepsis: Overview of the diagnostic and treatment challenges. Microbiol Spectr, 2015; 3:1–18.

14.

Singer

, Deutschman

, Warren Seymour

, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA, 2016; 315:801–810.

15.

Kozyrakis

, Perikleous

, Chatzistamou

, et al. Is there a role for Double J stent culture in contemporary urology?. Urol Int, 2018; 100:203–208.

16.

Borofsky

, Walter

, Shah

, Goldfarb

, Mues

, Makarov

. Surgical decompression is associated with decreased mortality in patients with sepsis and ureteral calculi. J Urol, 2013; 189:946–951.

17.

Blackwell

, Barton

, Kothari

, Zapf

, Flanigan

, Kuo

. Early intervention during acute stone admissions: Revealing “the weekend effect” in urological practice. J Urol, 2016; 196:124–130.

18.

Wolf

. Percutaneous approach to the upper urinary tract collecting system, Chapter 47. In: Wein

, Kavoussi

, Novic

, Partin

, Peters

, eds. Campbell-Walls' Urology, 10th Edition. Philadelphia, PA: Elsevier, 2012, pp. 1321–1356.

19.

Yossepowitch

, Lifshitz

, Dekel

, et al. Predicting the success of retrograde stenting for managing ureteral obstruction. J Urol, 2001; 166:1746–1749.

20.

Ghani

, Sammon

, Trinh

. The infected upper urinary tract stone and obstruction. Stent first where possible?. Eur Urol, 2013; 64:93–96.

21.

Ramchandani

, Cardella

, Grassi

, et al. Quality improvement guidelines for percutaneous nephrostomy. J Vasc Interv Radiol, 2003; 14(9 Pt 2):S277–S281.

22.

Matlaga

How do we manage infecred obstructive pyelonephritis?. Eur Urol, 2013; 64:93–94.

23.

Oken

, Creech

, Tormey

, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol, 1982; 5:649–655.

24.

Kelly

, Shahrokni

. Moving beyond Karnofsky and ECOG performance status assessments with new technologies. J Oncol, 2016; 2016:6186543.

25.

Schag

, Heinrich

, Ganz

. Karnofsky performance status revisited: Reliability, validity, and guidelines. J Clin Oncol, 1984; 2:187–193.

26.

Charlson

, Szatrowski

, Peterson

, Gold

. Validation of a combined comorbidity index. J Clin Epidemiol, 1994; 47:1245–1251.

27.

Chang

, Yin

, Wei

, et al. Adjusted age-adjusted Charlson comorbidity index score as a risk measure of perioperative mortality before cancer surgery. PLoS One, 2016; 11:e0148076.

28.

Huang

, Tsai

, Yu

, Ko

. Epidemiology and outcome of severe sepsis and septic shock in surgical intensive care units in Northern Taiwan. Medicine (Baltimore), 2015; 94:e2136.

29.

Wagenlehner

, Tandogdu

, Bjerklund Johansen

. An update on classification and management of urosepsis. Curr Opin Urol, 2017; 27:133–137.