emqfulsttfpiaomeo1sbsvIanI

ep

RA

The Journal of Emergency Medicine, Vol. 39, No. 2, pp. 198–203, 2010Copyright © 2010 Elsevier Inc.

Printed in the USA. All rights reserved0736-4679/$–see front matter

doi:10.1016/j.jemermed.2009.02.006

Ultrasound inEmergency Medicine

CAN BEDSIDE ULTRASOUND ASSIST IN DETERMINING WHETHER SERUMCREATININE IS ELEVATED IN CASES OF ACUTE URINARY RETENTION?

Kaushal Shah, MD,* Jennifer Teng, MD,* Hiral Shah, MD,† Alice Choe, MD,* Amir Darvish, MD,‡David Newman, MD,* and Dan Wiener, MD*

*Department of Emergency Medicine, St. Luke’s-Roosevelt Hospital, New York, New York, †Department of Emergency Medicine,North Shore University Hospital, New Hyde Park, New York, and ‡Department of Emergency Medicine, Tufts-New England Medical

Center, Boston, MassachusettsReprint Address: Kaushal Shah, MD, Department of Emergency Medicine, St. Luke’s-Roosevelt Hospital, 1111 Amsterdam Ave, New

York, NY 10025

ApAAm1lno

dwiTe

tpcnhp

e

Abstract—Background: There are no guidelines to deter-ine which patients with acute urinary retention (AUR) re-

uire blood testing (i.e., serum creatinine) to assess for renalailure. Objective: To determine if hydronephrosis on bedsideltrasound correlates with an abnormal serum creatinine (Cr)

evel in cases of AUR. Methods: This was a prospective, ob-ervational study of subjects clinically diagnosed with AUR atwo associated urban academic centers from October 2004hrough August 2006. Emergency physicians completed a dataorm and performed a bedside ultrasound to determine theresence or absence of hydronephrosis. The data collected

ncluded suspected cause of AUR, amount of urinary outputfter Foley insertion, and blood test results. Follow-up wasbtained by telephone and electronic medical record for 1onth. Standard statistics were employed. Results: Among 96

nrolled subjects with AUR, 43 had a serum Cr level obtainedn the initial visit, and 10 (23%; 95% confidence interval [CI]1–36) of these had an elevated Cr (10% [95% CI 4–16] of thetudy cohort). The test characteristics of hydronephrosis onedside ultrasound to detect elevation in Cr were a sensitivity,pecificity, positive predictive value, and negative predictivealue of 70%, 67%, 39%, and 88%, respectively. Conclusion:n cases of AUR, the prevalence of elevated creatinine is high,nd hydronephrosis based on bedside ultrasonography doesot correlate with elevation in creatinine. © 2010 Elseviernc.

Keywords—urinary retention; creatinine; ultrasonogra-hy; hydronephrosis

ECEIVED: 14 October 2008; FINAL SUBMISSION RECEIVED:

CCEPTED: 6 February 2009

198

INTRODUCTION

cute urinary retention (AUR) is a common medicalroblem in the emergency setting. Most patients withUR are men over the age of 60 years, and the risk ofUR increases significantly with age (1,2). It is esti-ated that at least one episode of AUR will develop in

0% and 33% of men in their 70s and 80s, respectively,argely due to this population’s high prevalence of be-ign prostatic hypertrophy, one of the most commonbstructive etiologies of AUR (2–4).

AUR is characterized by a sudden inability to urinateespite a distended bladder, and patients often presentith marked discomfort, difficulty voiding, and a phys-

cal examination revealing a palpable, enlarged bladder.reatment of simple AUR should focus on prompt cath-terization to decompress the bladder (5).

Beyond initial decompression of the bladder, the role ofhe emergency care provider in cases of AUR has beenoorly discussed and studied, both in textbooks of acuteare and in the peer-reviewed literature. Although hydro-ephrosis and renal damage may occur in such cases, thisas been a largely unreported phenomenon, making therevalence and clinical importance of these entities unclear.

The purpose of this study was to determine the pres-nce or absence of hydronephrosis on bedside ultrasound

cember 2008;

12 De

oio

S

Twjrp

S

Tuc2au

S

Btstkhac

ssm

apsEkiaf(ppi

btEtTp

D

SshainFtn

ottctrt

D

“tclpE

Fe

Elevated Serum Creatinine in Acute Urinary Retention 199

f the kidneys and to determine if ultrasonographic find-ngs correlate with an abnormal creatinine level in casesf AUR presenting to the Emergency Department (ED).

MATERIALS AND METHODS

tudy Design

his was a prospective, observational study. The studyas approved by our Institutional Review Board. Sub-

ects were identified by emergency physicians (EPs),esidents, and research assistants working in the ED. Allarticipants gave informed written consent.

tudy Setting and Participants

he study was conducted in the EDs at two associatedrban academic centers, with a combined annual patientensus of 150,000, from October 2004 through August006. Adult patients diagnosed clinically with AUR byn EP were eligible for enrollment in the study. Patientsnable to provide informed consent were excluded.

tudy Protocol

efore the start of the study, all EPs and residents wererained on the use of bedside ultrasound by the ultra-ound division of the ED (specifically, the ultrasound-rained EPs and the ultrasound fellows) to identify theidney in two views and the ultrasonographic findings ofydronephrosis. All physicians obtained individualizedttention with hands-on training with the ultrasound ma-hines to identify the kidneys in two planes.

Both EDs are staffed with research assistants trainedpecifically for identifying and enrolling patients in thetudy. Research assistant staffing in the ED is approxi-ately 16 h per day.When a patient with AUR was identified by research

ssistant or physician staff, consent was obtained and theatient was approached for enrollment in the study. Thetudy did not dictate any change in medical management.Ps were asked to perform a bedside ultrasound of bothidneys in two views (longitudinal and transverse) eithern the supine or decubitus position approximately 15 minfter Foley catheter insertion, and to complete a dataorm containing their interpretation of the ultrasoundi.e., absence or presence of hydronephrosis) and basicatient information. The completed data form was thenlaced in a sealed box that was emptied by the study

nvestigators on a regular basis. o

Although not required by the study protocol, a num-er of static ultrasound scan images were submitted withhe data form; these were reviewed by a board-certifiedP with ultrasound fellowship training and certification by

he Registry for Diagnostic Medical Sonography (RDMS).he RDMS-certified EP was blinded to the original inter-retation of the ultrasounds.

ata Collection

tandardized data forms were available at both hospitalites throughout the study period. Illustrative images ofydronephrosis were included in the data form to providevisual reminder to the EPs (Figure 1). Data collected

ncluded: past medical history, suspected cause of uri-ary retention, amount of urinary output 15 min afteroley insertion, results of laboratory studies (if ob-

ained), interpretation of presence or absence of hydro-ephrosis, final diagnosis, and disposition from the ED.

After the initial ED visit, 2–4-week follow-up wasbtained with telephone calls to the subject and review ofhe hospital record for creatinine levels before and afterhe initial visit. Standardized follow-up questions in-luded: whether they had a follow-up urology visit, ifhere were any problems since leaving the ED (i.e.,epeat hospital visit or repeat obstruction episode), and ifhey knew their urologic diagnosis.

efinitions

Urinary retention” was based on a clinical diagnosis byhe EP. “Hydronephrosis” is defined as a dilatation of theollecting system, represented on ultrasonography asarge, echo-free areas within the renal sinus, but theresence or absence of hydronephrosis was based on theP’s interpretation with only the aide of an illustration

igure 1. Images on data form to assist selection of pres-nce or absence of hydronephrosis (7).

n the data form (6,7). Elevated creatinine was defined as

g1

O

Tsevcntgt

D

WpcdA

bRwucs

N6nPb2(n(

5ocs[(hrw

phtt(inn

T

MFBHDPN

CH

T

A

200 K. Shah et al.

reater than the hospital laboratory normal range (above.3) or 0.5 above subject’s baseline creatinine.

utcomes

he main outcome measure was hydronephrosis on bed-ide ultrasonography and whether it correlated with anlevated creatinine level. Given that this was an obser-ational study and we did not require EPs to obtain areatinine level on all patients with AUR, we used aormal subsequent creatinine or no complications relatedo the initial visit within 2–4-week follow-up as surro-ate measures for no elevation in creatinine on the day ofhe ED visit for AUR.

ata Analysis

e used descriptive statistics to describe the subjectopulation. Standard statistical methods were used toalculate diagnostic test operating characteristics of hy-ronephrosis as a predictor of elevated serum creatinine.

kappa score was calculated to determine agreement

able 1. Characteristics of Cohort (Patients with AcuteUrinary Retention)

Characteristics n (%) [95% CI]

ale 85 (88.5%) [82.2–94.9]emale 11 (11.5%) [5.1–17.8]PH 55 (57.3%) [47.4–67.2]TN 22 (22.9%) [14.5–31.3]M 10 (10.4%) [4.3–16.5]rostate cancer 5 (5.2%) [0.8–9.7]o BPH, HTN, DM, prostate cancer 28 (29.2%) [20.1–38.3]

I � confidence interval; BPH � benign prostatic hypertrophy;TN � hypertension; DM � diabetes mellitus.

able 2. Acute Urinary Retention Subjects with Elevated Cr

Age (years) Sex Disposition Baseline Cr Visit Cr Su

70 M Admission 1.3 7.079 M Admission NA 6.776 M Left AMA NA 6.757 M Admission 1.5 4.769 M Admission NA 3.472 F Admission NA 2.972 M Admission NA 2.465 M Home NA 2.190 M Home 1.0 1.784 M Home 1.0 1.585 M Home 1.3 NT83 M Home NA NT

MA � against medical advice; NA � not available; NT � not tested;

etween EP determination of hydronephrosis and theDMS-certified EP. Spearman correlation coefficientsere used to analyze the two continuous variables ofrine output and level of creatinine. Categorical andontinuous variables were analyzed using the chi-quared test and the Mann-Whitney test, respectively.

RESULTS

inety-six subjects (11 female and 85 male; mean age8 � 16 years, range 16–99 years) with clinically diag-osed acute urinary retention were enrolled in the study.ast medical history of the cohort included: 55 (57.3%)enign prostatic hypertrophy, 5 (5.2%) prostate cancer,2 (22.9%) hypertension, 10 (10.4%) diabetes mellitus, 22.1%) chronic renal insufficiency, and 28 (29.2%) withone of these diagnoses (Table 1). Thirteen subjects13.5%) were lost to follow-up.

Forty-three (45%; 95% confidence interval [CI] 35–5]) subjects had blood tests performed at the discretionf the treating EP. Thirteen subjects had an elevatedreatinine, including 3 that were consistent with theubject’s baseline; therefore, 10 AUR subjects (23%95% CI 11-36] among blood-tested subjects) or 10%95% CI 4-16) of the study cohort were considered toave an elevation in creatinine. Elevated creatinine levelsanged from 1.5–7.0, with a mean of 3.9 � 2.2. No subjectas found to have an elevated serum potassium level.Among the subjects who did not have blood testing

erformed in the ED, 2 were subsequently confirmed toave an elevated creatinine within 2 weeks. Assuminghese subjects had an undiagnosed elevated creatinine onhe day of their ED visit for AUR, a total of 12 subjects12%; 95% CI 6–19) in our cohort had a new elevationn creatinine. At 4 weeks, 4 of these subjects had aormal creatinine, 6 had a persistently elevated creati-ine, and 2 subjects were lost to follow-up. Table 2

e (Cr)

nt Cr Cr Normal at 4 Weeks? Hydro Urine Output (mL)

0 N Y 13007 N Y 2000

NA Y 14008 N Y 6001 Y N 400

Y Y 10000 N N 9001 Y Y 2000

Y N 800NA Y 900N N 800N N 900

eatinin

bseque

3.0,2.4.1,1.

NA3.6,3.1.4,1.1.01.6,2.1.6,1.1.0

NA1.91.6

Hydro � hydronephrosis on bedside ultrasound.

sn

9rcnsaTh1

wCcssaThCwp0

msaaofw1wh

uwpeaEsaRia

Oc

pcvpw

Falcwta

ttovOrtecctpgotcntsGlej

asmwswih

srdapl

Elevated Serum Creatinine in Acute Urinary Retention 201

hows follow-up and ED visit data for subjects withewly elevated creatinine.

Of the 43 subjects with creatinine testing, 18 (41.9%;5% CI 27–57) had hydronephrosis on bedside ultrasonog-aphy. Among the 10 subjects with a new elevation inreatinine on the index visit, 7 had hydronephrosis and 3 didot, for a sensitivity of 70% (95% CI 42–98). Among the 33ubjects with a normal creatinine, 11 had hydronephrosisnd 22 did not, for a specificity of 67% (95% CI 51–83).he positive and negative predictive values of bedsideydronephrosis for elevated creatinine were 39% (95% CI6–61) and 88% (95% CI 75–100), respectively.

Of the complete cohort with AUR, hydronephrosisas sonographically present in 30 subjects (31.3%; 95%I 22–41). Among the 12 subjects with an elevatedreatinine, 7 had hydronephrosis and 5 did not, for aensitivity of 58% (95% CI 30–86). Among the 84ubjects with a normal creatinine, 23 had hydronephrosisnd 61 did not, for a specificity of 73% (95% CI 63–82).he positive and negative predictive value of bedsideydronephrosis for elevated creatinine were 23% (95%I 8–38) and 92% (95% CI 86–99), respectively,hereas the confidence intervals for both negative andositive likelihood ratios crossed unity at 1.5 (95% CI.7–3.2) and 0.8 (95% CI 0.5–1.3), respectively.

Urine output approximately 15 min after Foley place-ent was documented on 89 subjects. There was no

tatistically significant correlation between urine outputnd level of creatinine (r2 � 0.07, p � 0.65). In post hocnalysis, using 750 mL as a threshold for “high urineutput” (HUO) vs. “low urine output” (LUO), the dif-erence in number of subjects with elevated creatinineas statistically significant. Ten of 41 (24%; 95% CI1–38) HUO subjects had an elevated creatinine,hereas only 2 of 48 (4%; 95% CI 0–10) LUO subjectsad an elevated creatinine.

Thirty-eight data forms were submitted with staticltrasound images of the kidneys in two views. Theseere all over-read by an RDMS-certified EP for theresence or absence of hydronephrosis without knowl-dge of the original interpretation. There was “moderate”greement (kappa � 0.6; 95% CI 0.2–0.9) between theP and the RDMS-certified EP. Among the 38 initialonograms, hydronephrosis was identified in 9 subjects,nd absence of hydronephrosis was identified in 29. TheDMS-certified EP interpreted the printed ultrasound

mages and agreed on 4 of the 9 cases of hydronephrosisnd all 29 cases of absence of hydronephrosis.

DISCUSSION

ur study identified a surprising prevalence of elevated

reatinine in a convenience cohort of 96 individuals e

resenting with AUR. The prevalence of an elevatedreatinine among those subjects tested on the day of theirisit was 23.3%. Similarly, the number of subjects withersistent, new elevations in creatinine at 1 month (6.7%)as also higher than expected.The general assumption is that AUR is corrected with

oley catheterization and therefore, it is not clear whetherny laboratory tests are necessary. There are no clear guide-ines. Blood testing is done at the discretion of the physi-ian. We were unable to ascertain from our study designhy serum tests were ordered, but these pilot data suggest

hat EPs may consider blood testing more frequently whilewaiting a larger, more rigorous study of AUR.

Given the convenience nature of our sample, it isempting to consider attributing our findings to a selec-ion bias. Given the methods utilized in this and multiplether formal, Institutional Review Board-approved in-estigations in our ED, this explanation seems unlikely.ur trained research assistants initiate virtually all en-

ollments using access to a computer-based tracking sys-em and chief complaint-driven triggers for potentialnrollment. For enrollment in this investigation, physi-ians were asked to perform ultrasound examination andomplete a data form describing the ultrasound and pa-ient characteristics. Given the potential minor delays inatient care, including decompression of the bladder, andiven the potential for provider or patient refusal and theccasional inability of research assistants to identify po-ential subjects in a timely fashion (before bladder de-ompression, which is occasionally undertaken by ourursing staff without physician involvement), both ofhese factors seem likely to skew the cohort toward a lessevere, less distressed, less compromised population.iven these conditions and the lack of previously pub-

ished data from this patient population, the finding of anlevated creatinine in approximately 1 of every 8 sub-ects is cause for concern.

Alternatively, perhaps a lack of subjective distress orpparent discomfort is a marker for patients with longer-tanding urinary retention and therefore, renal impair-ent associated with their AUR event. This possibilityill require more formal investigation. In any case, it

eems clear that until further data suggest that our cohortas either non-representative or statistically aberrant, it

s reasonable to presume that those presenting with AURave a significant chance of creatinine elevation.

One goal of this investigation was to assess the fea-ibility and potential utility of bedside renal ultrasonog-aphy in the setting of AUR. This seems an ideal mo-ality to study and image the kidneys secondary to theccessibility of the kidneys and the different acousticroperties present in the renal cortex, medulla, and col-ecting system (6). We hypothesized that if creatinine is

levated from baseline as a result of AUR, the patient

sarmwaapt

otacGlon

nimpmsipnvbuse

wupfccs

Ombnoms

t

jhqon

psrs

fcc

prdptmpp

Abhis

Itdct3

(pit

1

2

3

202 K. Shah et al.

hould have hydronephrosis on bedside ultrasound, andn absence of hydronephrosis would suggest a milderetention episode and no elevation in creatinine. Theajor intent of our study, therefore, was to determinehether the presence of hydronephrosis correlates with

bnormalities in renal function in cases of AUR. Wessumed that EPs could successfully identify hydrone-hrosis on bedside ultrasound after a brief in-serviceraining.

In our study, absence or presence of hydronephrosisn bedside ultrasound did not correlate well with eleva-ion in creatinine. A sensitivity and specificity of 70%nd 67%, respectively, are too low to use hydronephrosislinically as a test to determine the need for blood testing.iven our small study population, it is still possible that a

arger study using well-trained ultrasonographers may dem-nstrate that ultrasound can distinguish which AUR patientseed serum creatinine testing.

The ability of EPs and residents to accurately diag-ose hydronephrosis (after only a single in-service train-ng) is questionable given the only “moderate” agree-ent with the RDMS-certified ultrasonographer. One

revious Emergency Department study demonstrateduch higher test characteristics of sensitivity (94%) and

pecificity (96%) to rule out hydronephrosis; however,mages were recorded on videotape for review (8). It isossible that the static printed images in our study didot demonstrate hydronephrosis as well as the real-timeideo images. Another study of bedside ultrasonographyy EPs for hydronephrosis in patients with suspectedreteral colic demonstrated a sensitivity of 72% andpecificity of 73% using computed tomography and py-lography as the standard (9).

Of additional interest is the correlation of urine outputith elevation in creatinine. Based on post hoc analysis,rine output approximately 15 min after Foley catheterlacement may reflect the severity of retention and the needor blood testing. The majority of patients with an elevatedreatinine (10 of 12) had a urine output above 750 mL. Theutoff of 750 mL was not determined a priori and thereforehould be further tested prospectively.

LIMITATIONS

urs is a convenience sample. This is mitigated by ourethodology as much as possible, but certain selection

iases are likely to affect these data. As with all conve-ience samples, understanding the specific methodologyf our data collection may help one to understand theanner in which validity will be impacted in our selected

ample of AUR patients.This was an observational study and, as a result, blood

esting results were not available for many of the sub-4

ects. The prevalence of elevated creatinine may be evenigher than determined in this study. We also did notuery the physicians on why they chose to obtain or notbtain blood testing on their patients. A baseline creati-ine was also not available for all of the subjects.

Although not demonstrated by this study, it is stillossible that bedside ultrasound may play a role intratifying which patients warrant blood testing for se-um creatinine. A larger study using well-trained ultra-onographers may prove this to be the case.

Thirteen and a half percent of subjects were lost toollow-up. It is possible that these subjects had compli-ations after their initial visit (i.e., persistently elevatedreatinine).

There are no hard criteria for diagnosing hydrone-hrosis based on bedside ultrasound. Dilatation of theenal pelvis is a subjective determination that may be tooifficult after only a brief in-service and little-to-no ex-erience with cases of definite hydronephrosis. In addi-ion, hydronephrosis may be masked by dehydration orimicked by some other common processes, including

regnancy, renal cortical cysts, or prominent medullaryyramids (anatomical variant).

Our study did not inquire about the time duration ofUR before Foley placement. It is possible that this maye a factor (in addition to amount of urine output) thatelps distinguish which patients are at risk for elevationn creatinine. It should also be documented in futuretudies.

CONCLUSIONS

n cases of AUR, hydronephrosis based on bedside ul-rasonography by emergency physicians and residentsoes not correlate with elevation in creatinine. The testharacteristics of sensitivity, specificity, positive predic-ive value, and negative predictive value are 70%, 67%,9%, and 88%, respectively.

Given the high prevalence of an elevated creatinineat least 10–12%) in this cohort of subjects, emergencyhysicians interested in identifying creatinine elevationn patients with AUR should maintain a low threshold foresting the serum creatinine.

REFERENCES

. Emberton M, Anson K. Acute urinary retention in men: an age oldproblem. BMJ 1999;318:921–5.

. Handrigan MT, Thompson I, Foster M. Diagnostic procedures forthe urogenital system. Emerg Med Clin North Am 2001;19:745–61.

. Lepor H. Managing and preventing acute urinary retention. RevUrol 2005;7:S26–33.

. Curtis L, Dolan T. Acute urinary retention and urinary incontinence.Emerg Med Clin North Am 2001;19:591–619.

5

6

7

8

9

Elevated Serum Creatinine in Acute Urinary Retention 203

. Rosenstein D, McAninch J. Urologic emergencies. Med Clin NorthAm 2004;88:495–518.

. O’Neill WC. Sonographic evaluation of renal failure. Am J KidneyDis 2000;35:1021–39.

. Swadron S, Mandavia D. Renal. In: Ma OJ, Matteer J, eds. Emer-

gency ultrasound. New York: McGraw-Hill; 2003:197–220.

. Lanoix R. Preliminary evaluation of emergency ultrasound in thesetting of an emergency medicine residency training program. Am JEmerg Med 2000;18:41–5.

. Rosen CL, Brown DF, Sagarin MJ, Chang Y, McCabe CJ, WolfeRE. Ultrasonography by emergency physicians in patients with

suspected ureteral colic. J Emerg Med 1998;16:865–70.

ARTICLE SUMMARY1. Why is this topic important?

There are no guidelines to determine which patientswith acute urinary retention require blood testing (i.e.,serum creatinine) to assess for renal failure.2. What does this study attempt to show?

Hydronephrosis diagnosed in the emergency depart-ment by bedside ultrasonography of the kidneys canpredict which patients will have an elevation in creati-nine.3. What are the key findings?

The prevalence of an elevated creatinine was high(approximately 12%), but bedside ultrasonography forhydronephrosis could not predict which subjects had anelevation.4. How is patient care impacted?

Emergency physicians interested in identifying creati-nine elevation in patients with acute renal failure shouldmaintain a low threshold for testing the serum creatinine.