Article, Urology

Urinary obstruction is an important complicating factor in patients with septic shock due to urinary infection

a b s t r a c t

Objective: urinary tract infection is a common cause of severe sepsis, and anatomic urologic obstruction is a recognized factor for complicated disease. We aimed to identify the incidence of Urinary obstruction complicat- ing acute septic shock and determine the characteristics and outcomes of this group.

Methods: Patients prospectively enrolled in a sepsis treatment pathway registry between October 2013 and July 2014 were reviewed for the diagnosis of UTI. Standardized medical record review was performed to confirm sep- sis due to UTI and determine clinical variables including the presence of anatomic urinary obstruction. Patients with septic shock due to UTI with obstruction were compared with those without obstruction. The primary out- comes were incidence of urinary obstruction and hospital mortality.

Results: Among 1084 registry enrollees, 209 (19.2%) met inclusion criteria for the study. Acute anatomic obstruc- tion was identified in 22 (10.5%) patients. Hospital mortality in patients with obstruction was 27.3% compared with 11.2% in patients without obstruction (absolute difference of 16.1%; P = .03; 95% confidence interval [CI], 1.2%-30.9%). Hospital length of stay among survivors was 12.8 days compared with 8.3 days (absolute difference of 4.5 days; P = .04; 95% CI, 0.2-8.8 days). History of urinary stone disease was independently associated with obstruction (odds ratio, 5.6; 95% CI, 2.2-14.3).

Conclusions: Approximately 1 in 10 patients presenting with septic shock due to a urinary source is complicated by anatomic urinary obstruction. These patients have significantly higher mortality compared with patients without obstruction. Early imaging of patients with septic shock due to suspected urinary source should be con- sidered to identify obstruction requiring emergency intervention.

Introduction

Approximately 750,000 cases of sepsis occur annually in the United States, and the incidence continues to rise [1,2]. Infections of the urinary tract (UTIs) account for approximately 20% of severe sepsis cases [3,4]. Anatomic obstruction of the urinary tract is a recognized source of com- plicated UTI. Contemporary septic shock and severe sepsis management focuses on early recognition, aggressive fluid resuscitation, hemody- namic support, and infection source control [1,5,6]. Early urologic inter- vention for patients with UTI complicated by obstruction is a recognized means of source control, but the incidence of this complicating factor and recommendations for imaging and management are not current- ly defined. This study was performed to identify the incidence of uri- nary obstruction complicating septic shock due to a urinary source

E-mail addresses: [email protected], [email protected] (A.C. Heffner).

and determine the outcomes and characteristics of this group com- pared with patients without obstruction.

Methods

Design and setting

Carolinas Healthcare System (CHS) uses a dedicated treatment path- way for high-risk patients with acute severe sepsis. Enrollment in the pathway requires suspicion of infection and systolic blood pressure b 90 mm Hg after 20 mL/kg of intravenous fluids or serum lactate

>= 4.0 mmol/L. All patients enrolled in this clinical pathway undergo

protocol-based quantitative resuscitation including early antimicrobial administration. A Code Sepsis patient registry is maintained for quality improvement purposes. All study patients were admitted through the emergency department (ED) at one of several urban hospitals in the Carolinas Healthcare Network within metropolitan Charlotte, NC. The Institutional Review Board and Privacy Board of Carolinas Healthcare System approved this study under waiver of informed consent.

http://dx.doi.org/10.1016/j.ajem.2015.12.068

Identification of subjects

K. Reyner et al. / American Journal of Emergency Medicine 34 (2016) 694–696

Table 1

Study patient characteristics by study group

695

All adult patients prospectively enrolled in the Code Sepsis clinical pathway between October 2013 and July 2014 with an International Classification of Diseases, Ninth Revision, diagnosis of UTI were eligible. Standardized medical record review was performed on all of these cases. Subjects were included if they were N 18 years of age, were en- rolled in the Code Sepsis pathway within 24 hours of ED triage, and had clinically confirmed septic shock due to a urinary source. Patients were excluded if they had an alternative diagnosis or source of sepsis. UTI was defined by pyuria (N 5 white blood cells per high-power field) on urinalysis or culture-confirmed uropathogen in the appropriate Clinical context.

Variable Total group

(N = 209)

Age (y)	70 +- 15	67 +- 18	70 +- 14	.42
Sex (male)	97 (46%)	8 (36%)	89 (48%)	.32
Comorbidities
Diabetes	73 (35%)	6 (27%)	67 (36%)	.43
Hypertension	145 (69%)	11 (50%)	134 (72%)	.04
Coronary disease	50 (24%)	4 (18%)	46 (25%)	.51
Heart failure	37 (18%)	4 (18%)	33 (18%)	.95
CKD	43 (21%)	4 (18%)	39 (21%)	.77
ESLD	14 (7%)	0 (0%)	14 (7%)	.18
ESRD	1 (0.5%)	0 (0%)	1 (0.5%)	.73
History of urinary stone	44 (21%)	12 (55%)	32 (17%)	b.001
Hospitalization within 60 d Clinical variables	68 (33%)	5 (23%)	63 (34%)	.30
Triage temp (?F)	99.1 +- 2.8	99.6 +- 2.4	99.1 +- 2.9	.49
ED fever	87 (42%)	10 (45%)	77 (41%)	.70
ED hypothermia	28 (13%)	0 (0%)	28 (15%)	.05
Triage HR	106 +- 27	107 +- 23	106 +- 27	.87
Triage SBP	107 +- 27	107 +- 31	107 +- 27	.90
Lowest ED SBP	85 +- 20	83 +- 23	86 +- 20	.60
TriAge SI	1.1 +- 0.4	1.1 +- 0.4	1.1 +- 0.4	.64
1st lactate	4.2 +- 2.8	4.4 +- 2.6	4.2 +- 2.8	.75
Max ED lactate	4.2 +- 2.8	4.4 +- 2.6	4.2 +- 2.8	.73
Admission AKI^a	123 (67%)	12 (80%)	111 (66%)	.27
Acute imaging	88 (42%)	22 (100%)	66 (35%)	b.001
Urine culture positive	181 (87%)	18 (82)%	163 (86%)	.51
Bacteremia	75 (36%)	14 (64%)	61 (33%)	b.01

Obstruction (n = 22)

Nonobstruction

(n = 187) P

Data collection and analysis

A standardized data collection instrument and guidance tool were developed for data collection. Record review and data abstraction were performed by a single author. A second blinded reviewer assessed equivocal cases for classification. Patients were divided into 1 of 2 groups according to the presence or absence of anatomic urinary ob- struction. Obstruction was defined by the presence of structural obstruc- tion with hydronephrosis by radiologic imaging. Admission acute kidney injury was defined by creatinine rise greater than 0.3 mg/dL from base- line. Continuous data are presented as means +- standard deviation or medians and interquartile ranges and, when appropriate, were tested for statistical differences using unpaired t tests or Mann-Whitney U tests. Categorical data are reported as counts and percentages and were tested for significance using ?² or Fisher exact tests when applica- ble. For all statistical 2-tailed tests, P b .05 was considered significant.

To identify factors independently associated with urinary obstruc-

tion complicating septic shock due to UTI, we fit a multivariate logistic regression model using obstruction as the dependent variable. Candi- date variables were selected based on significant associations (P b .1) on univariate analysis. The model was refined using backward stepwise elimination. model fit was assessed by standardized Pearson test. Odds ratios (ORs) and confidence intervals (CIs) are presented for the final model.

Results

During the study period, 1084 patients were entered in the Code Sepsis pathway, of which 311 (29%; 95% CI, 26%-31%) patients had an In- ternational Classification of Diseases, Ninth Revision, diagnosis of UTI. One hundred and two patients were excluded because they had an alterna- tive diagnosis or source of infection, leaving 209 (19.3% of master regis- try group; 95% CI, 17.0-21.7) patients for study analysis. The demographic and clinical characteristics of the study subjects are re- ported in Table 1. ED clinical variables including hemodynamics, lactate, and acute organ dysfunction did not differ between the 2 groups. Pa- tients with obstruction were more likely to undergo radiologic imaging, have history of urinary stone disease, and have confirmed bacteremia based on admission blood cultures.

Anatomic urinary obstruction was identified in 22 (11%; 95% CI, 16%- 27%) patients. Sources of obstruction are detailed in Table 2. Ureteral ob- struction was most common and was identified in 17 (77%; 95% CI, 56%- 90%) patients. Acute intervention for urinary decompression was per- formed in 19 patients (86%). One patient died before intervention, and 2 patients had directives precluding intervention.

Hospital mortality in patients with obstruction was 27.3% compared with 11.2% in patients without obstruction (absolute difference of 16.1%, P = .03; 95% CI, 1.2%-30.9%). Hospital length of stay among sur- vivors was 12.8 days compared with 8.3 days (absolute difference of 4.5 days, P = .04; 95% CI, 0.2-8.8 days) (Table 3).

Multivariate logistic regression analysis identified that history of uri- nary stone disease and bacteremia identified via admission blood

Abbreviations: CKD, chronic kidney disease; ESLD, end-stage liver disease; ESRD, end-stage renal disease; SBP, systolic blood pressure in mm Hg; HR, heart rate in beats per minute; SI, Shock Index (HR/SBP); AKI, acute kidney injury.

^a Acute kidney injury among patients with baseline.

culture were independently associated with urinary obstruction. Devi- ance and Pearson goodness-of-fit statistics were equivalent at a test sta- tistic of 0.61 (P = .43), indicating satisfactory model fit. Odds of urinary obstruction were 5.6 times more likely with history of urinary stone dis- ease (OR 95% CI, 2.2-14.1) and 3.4 times as likely when bacteremia was identified (OR 95% CI, 1.3-8.9). Area under the curve of the receiver op- erating characteristic curve for the model was 0.77.

Discussion

Our results demonstrate that anatomic urinary obstruction is a rela- tively common complication of severe sepsis secondary to a urinary source. Compared with patients without obstruction, these patients had significantly higher morbidity and mortality. Given a limited body of existing literature focusing on this topic, our results provide some in- sight into this important complication.

The incidence of acute urinary obstruction complicating severe sep- sis in patients with UTI has not formerly been reported. Our work sug- gests that urinary obstruction complicating septic shock due to UTI is fairly common at a rate of 1 in 10. A former study of patients with acute bacteremic pyelonephritis admitted via a Korean ED identified urinary obstruction in 15% of patients and in 38% of patients presenting with shock [2]. The combination of these 2 investigations highlights a clinically important rate of anatomic obstruction among 2 patient pop- ulations presenting with septic shock.

Table 2

Etiology of urinary obstruction

Source of obstruction % of patients

Ureterolithiasis 68%

renal pelvis calculi 9%

Ureteral compression 9%

Bladder outlet obstruction 14%

696 K. Reyner et al. / American Journal of Emergency Medicine 34 (2016) 694–696

Table 3

Clinical outcomes of study groups

Variable	Total group	Obstruction	Nonobstruction	P
Death	27 (13%)	6 (27%)	21 (11%)	.03
Hospital LOS	8.1 +- 8.1	10.7 +- 9.4	7.8 +- 7.9	.12
Hospital LOS among survivors	8.7 +- 8.4	12.8 +- 10.2	8.3 +- 8.1	.04

Abbreviation: LOS, length of stay.

Our data also show that urinary infection in the context of urinary obstruction is a high-risk clinical situation. Patients with obstruction had significantly higher mortality and morbidity, as identified by hospi- tal length of stay. The former Korean pyelonephritis study also found urinary obstruction to be the strongest independent risk factor for septic shock (OR, 4.4) [7]. Unfortunately, our data set revealed few clinical markers to discriminate patients with obstruction at presentation. Al- though patients with obstruction were more likely to have history of urinary stone disease, this feature was not reliable to exclude obstruc- tion. The absence of readily available markers to identify obstruction is an important finding that supports early routine screening of patients presenting with septic shock due to presumed UTI.

Our work suggests that urinary obstruction complicating septic shock due to UTI is common enough to warrant routine early investiga- tion. Early recognition of obstruction is a critical step in source control. Sonographic evaluation of the urinary tract to rule out Obstructive uropathy in septic urinary patients has been previously recommended, but current guidelines do not focus on this point [8]. Increased use of Point-of-care ultrasonography makes it an appealing, rapid, and nonin- vasive means of initial screening, although reliability of ultrasound in the context of infection remains untested [9]. We endorse early imaging to evaluate for hydronephrosis with the goal of early intervention to achieve source control in identified cases. Identification of bladder out- let obstruction in a small number of cases represents an important high- light because it implies that obstruction was not identified based on clinical grounds before imaging. Despite increasing focus on avoidance of health care-associated UTI, we believe that urinary catheter insertion is warranted to establish urinary decompression and monitor urine out- put in critically ill patients with severe sepsis due to UTI.

Early source control is a cornerstone in the management of severe sepsis. The impact of timely antibiotics in severe sepsis is well recog- nized [10,11]. Appropriate source control also includes identification and decompression of infection sources based on the principles of drainage, debridement, device removal, decompression, and restoration of anatomy and function [12]. The international guidelines for the man- agement of severe sepsis recommend emergency source control within 12 hours of recognition (grade IC recommendation) [5]. Early source control is prioritized, based on evidence of efficacy, in common infection sources including complicated skin and soft tissue infections and infect- ed endovascular hardware [6]. The differential mortality of our group with urinary obstruction supports emergency urological interven- tion in these patients to achieve definitive source control. Decom- pression of the urinary tract can be accomplished via urinary catheter, ureteral stent, percutaneous nephrostomy, or, rarely, ne- phrectomy [13]. These findings also support further investigation to ascertain the window of opportunity to improve outcomes via rapid intervention.

Limitations

There are several limitations to our study. The most important of these is the retrospective design, which is associated with potential biases. To address these possible biases, we had a second blinded reviewer assess equivocal cases for classification. Although variables were extracted from a prospectively collected database, our results need confirmation in a prospective clinical study. A larger trial is also warranted given the small sample of obstructed patients in our analysis. Urinary imaging was not systematically performed on patients, which may have under- estimated the true incidence of obstruction in this cohort. Similarly, it is unclear how many patients with bladder outlet obstruction were identi- fied clinically and treated. Our database measured mortality in the hospi- tal, and longer follow up may have revealed different outcome rates.

Conclusions

Our study reveals that approximately 1 in 10 patients admitted to the hospital with septic shock due to UTI is complicated by anatomic urinary obstruction. These patients have significantly higher mortality compared with patients without obstruction. Early imaging of patients with acute septic shock due to suspected urinary source should be con- sidered to identify and characterize anatomic abnormalities that require emergency intervention.

Acknowledgments

Dr Heffner has received speaking honoraria from Edwards Lifesciences for presentations on topics related to sepsis. The authors have no additional affiliations or potential conflicts of interest to report. There were no sponsors or funding sources for this research.

References

AJEM