Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Epidemiology & Biostatistics

First Advisor

Ebbing Lautenbach


Coincident with the increasing use of fluoroquinolones (FQs) as the first-line agent for treatment of urinary tract infections (UTIs) in adults, the prevalence of high MIC fluoroquinolone susceptible E. coli (high MIC-FQSEC) which are the E. coli isolates with reduced susceptibility to FQs has increased substantially. The high MIC-FQSEC strains may serve as an important reservoir for FQ resistance in that treatment of these organisms with a FQ has been associated with future emergence of resistance.

To establish an effective program for controlling emergence of FQ resistance, it is necessary to understand the risk factors for, and impact of infection caused by high MIC-FQSEC. To identify risk factors for high MIC-FQ susceptibility, we conducted a case-control study of female subjects with UTIs caused by FQSEC at outpatient services within University of Pennsylvania Health System, Philadelphia. A total of 1836 subjects with low MIC-FQSEC UTI (CASE) and 165 subjects with high MIC-FQSEC UTIs (CONTROL) were enrolled into our study. Independent risk factors for high MIC-FQ included Asian race, having renal diseases and previous exposure to nitrofurantoin.

To determine the impact of high MIC-FQ susceptibility, we conducted a retrospective cohort study of female subjects with ambulatory FQSEC UTIs who were treated with FQ therapy. We enrolled 246 subjects into the low MIC (unexposed) group and 29 subjects into the high MIC (exposed) group. Study subjects with high MIC-FQSEC-UTIs were approximately 8 times more likely to experience treatment failure when received FQ therapy when comparing to those with low MIC FQSEC-UTIs.

The last dissertation project was a simulation study aiming to quantitatively compare the conventional case-control (CC) approach and the novel case-case-control (CCC) approach in investigating risk factors for infection caused by FQ-resistant pathogen. Our study confirmed that the CC approach almost always overestimates the effect of previous antibiotic exposure. The difference is more pronounced if the study is to be conducted among healthy population with a lower rate of colonization and protective effect of exposure on mechanism of harboring FQ-susceptible pathogen does not exist.

Additional Files

supplementarytable1.xls (238 kB)
supplementarytable2.xls (112 kB)