Warren Perry

Date of Award

January 2013

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Christopher L. Moore

Subject Area(s)



In theory, the evaluation of acute pyelonephritis (APN) does not require imaging, but in practice computed tomography (CT) may be ordered because of fear of an obstructing ureteral calculus that may present with similar symptoms. Childbearing age women are susceptible to the radiation exposure of CT imaging and have the highest incidence for APN. It is unknown what combination of clinical signs and point-of-care tests may help identify which patients with APN may require (or be able to avoid) advanced imaging.

We aimed to identify factors associated with ureteral stones in APN patients with the future goal of designing a decision rule that can identify these patients. We hypothesized that a set of clinical characteristics, including identification of hydronephrosis (which may be accomplished with point-of-care bedside ultrasound) would be able to differentiate complicated from uncomplicated APN.

This was a retrospective study of patients at an urban level-1 trauma ED and a freestanding 24-hr suburban ED. All CT-flank pain protocol (CTFPP) scans for renal colic between April 2005 and April 2009 were identified. We identified patients with APN as having >5 WBC/HPF on formal urinalysis in addition to one or more of the following: flank pain, CVA tenderness, chills, fever, nausea, or vomiting. Clinical data from the medical record were abstracted blinded to CT reports. Classification and regression tree analysis was used to produce decision trees and logistic regression was used to determine odds ratios.

In our study, 250 patients were included. Women composed 68.0% of the sample and the mean age was 41.8 (±15.6). The decision tree produced showed hydronephrosis to be the most predicting factor for a ureteral stones and the logistic regression also found a high statistical significant association with hydronephrosis, OR=29.03.

Our study is the first we are aware of to show that hydronephrosis is a dominant factor for predicting APN complicated by our ureteral stone. Our study also produced 2 clinically relevant decision tree that included hydronephrosis as a key finding for identifying patients with ureteral stones. Ultrasound can be used to detect hydronephrosis but there are few studies about its use in APN. Our findings show us the potential use in APN complicated by ureteral stones and support the further investigation of ultrasound for imaging decisions in cases of APN.


This is an Open Access Thesis.

Open Access

This Article is Open Access