Bacteremia and Mortality with Urinary Catheter–Associated Bacteriuria
Objective. Although catheter-associated urinary tract infection (CAUTI) and catheter-associated asymptomatic bacteriuria (CAABU) are clinically distinct conditions, most literature describing the risks of bacteriuria does not distinguish between them. We studied the relationship between catheter-associated bacteriuria and bacteremia from a urinary source in CAUTI relative to that in CAABU. Second, we investigated whether the presence or absence of urinary symptoms in catheterized patients with bacteriuria was associated with bacteremia from any source or mortality. Finally, we explored the effect of antimicrobial treatment of bacteriuria on subsequent bacteremia from any source and mortality.
Design. We performed a retrospective cohort study with 30 days of follow-up after an initial positive urine culture. CAUTI and CAABU were defined by Infectious Diseases Society of America guidelines.
Setting. A large tertiary care facility.
Patients. All inpatients with a urinary catheter (external or indwelling) and a positive urine culture between October 2010 and June 2011.
Results. We captured 444 episodes of catheter-associated bacteriuria in 308 patients; 128 (41.6%) patients had CAUTI, and 180 (58.4%) had CAABU. Three episodes of bacteriuria were followed by bacteremia from a urinary source (0.7%). CAUTI, rather than CAABU, was associated with bacteremia from any source, but neither CAUTI nor CAABU predicted subsequent mortality. Use of antimicrobial agents to treat bacteriuria was not associated with either bacteremia from any source or mortality.
Conclusions. Bacteremia from a urinary source was infrequent, and there was no evidence of an association of mortality with symptomatic versus asymptomatic bacteriuria in this population. Antibiotic treatment of bacteriuria did not affect outcomes.
Catheter-associated bacteriuria is extremely common worldwide,1 and it is estimated that more than 30 million indwelling catheters are placed annually in the United States, resulting in several hundred thousand cases of catheter-associated bacteriuria.2 Potential complications of catheter-associated bacteriuria include catheter-associated urinary tract infection (CAUTI), urinary-source bacteremia, and mortality.
Existing evidence suggests that catheter-associated bacteriuria is associated with a low rate of subsequent bacteremia and may not have a strong causal relationship with mortality. For example, in a prospective study of 1,497 patients with newly placed urinary catheters, 235 (15.7%) developed new-onset catheter-associated bacteriuria, but only 1 (0.07%) of these episodes clearly resulted in bacteremia.3 The relationship between urinary catheter–associated bacteriuria and mortality is more controversial. Platt et al,4 in 1982, attributed a 3-fold increase in mortality to catheter-associated bacteriuria. However, subsequent studies using a more standardized adjustment for severity of illness have not supported a causal relationship.5,6
Most literature on the clinical consequences of catheter-associated bacteriuria does not distinguish whether the patient had symptoms related to the urinary tract (ie, CAUTI) or whether the bacteriuria was asymptomatic (catheter-associated asymptomatic bacteriuria, or CAABU). Guidelines from the Infectious Diseases Society of America (IDSA) and other organizations have clearly distinguished between these two conditions and strongly advocate against treating CAABU.7-12 However, with the existing research gap, we do not know whether untreated CAUTI is more likely than CAABU to lead to negative outcomes in terms of bacteremia and mortality. Furthermore, we do not yet know whether treatment of CAUTI confers benefits beyond symptom relief, in terms of preventing bacteremia or mortality. Certain patient subgroups with catheter-associated bacteriuria may be at increased risk of bacteremia and thus be good candidates for targeted antimicrobial treatment.13-15 The idea of restricting antibiotic use to patients who are likely to benefit is a core tenet of the US Centers for Disease Control and Prevention’s (CDC’s) “Use Wisely” campaign to reduce antibiotic overuse.
We had 3 objectives. First, we sought to describe the relationship between catheter-associated bacteriuria and bacteremia from a urinary source, in CAUTI relative to CAABU. Second, we investigated whether the presence or absence of guidelines-defined urinary symptoms (CAUTI or CAABU) in catheterized patients with bacteriuria was associated with either bacteremia or mortality. Finally, we explored the effect of antimicrobial treatment of bacteriuria on bacteremia and mortality.
Setting, Design, and Study Population
We conducted a retrospective cohort study in 5 medicine and 5 extended care wards of a tertiary care facility, with approval from the Baylor College of Medicine Institutional Review Board and the Houston Veterans Affairs (VA) Research and Development Committee. Our extended care wards are in the same building as the acute care wards and encompass long-term assisted care as well as subacute rehabilitation beds. All patients who had both a urinary catheter and a positive urine culture (defined as a culture showing bacterial or fungal growth) between October 2010 and June 2011 were included. Our microbiology laboratory threshold for reporting bacterial or fungal growth in the urine is at least 103 organisms/mL, which meets the IDSA definition of CAUTI. We used the microbiology laboratory threshold for positive urine cultures at our institution, as this threshold reflects the clinician’s perspective. Both external (condom) catheters and indwelling urinary catheters meet the IDSA definition of urinary catheter use and were thus included in this study.
Classification of Bacteriuria as CAUTI or CAABU
Data were abstracted from the electronic medical records system by trained research personnel using an established chart-based surveillance method validated against a criterion standard of direct observation.16 Episodes of catheter-associated bacteriuria were classified by 2 investigators as either CAUTI or CAABU after a thorough systematic chart review starting from 2 days before to 1 week after the urine culture sample was collected, using a diagnostic pathway developed from the IDSA practice guidelines for CAUTI.7,17,18 We defined CAUTI as the presence of at least 103 colony-forming units/mL of at least one organism from the urine of a patient with one or more of the following signs and symptoms and no other identifiable cause: urgency, frequency, dysuria, suprapubic tenderness, pelvic discomfort, flank pain, rigors, gross hematuria, delirium, or fever (temperature of at least 100°F). To meet the definition of CAUTI, the patient had to be currently catheterized or to have had the urinary catheter removed within the preceding 48 hours. Frequency, urgency, and dysuria applied only to those patients whose catheter had been removed. We defined CAABU as bacteriuria or candiduria without any of these symptoms or as bacteriuria or candiduria with symptoms clearly attributable to a nonurinary cause.9 The rationale for using the clinical IDSA criteria rather than the surveillance CDC/NHSN (National Healthcare Safety Network) criteria was that the IDSA guidelines are written from the clinician’s perspective and focus on management decisions, while the CDC/NHSN criteria are intended for surveillance rather than to guide clinical decision making. The kappa statistic for the level of agreement between members of the research team regarding classification of cases of catheter-associated bacteriuria as CAABU versus CAUTI was 0.73 (86.7% agreement), or substantial agreement.19,20 All urine samples cultured within 7 days of the collection of the original positive culture sample were considered part of the same episode.
Detecting Bacteremia and Mortality
All positive blood cultures and all deaths from 1 day before to 30 days after the urine culture (defined as the time the sample for the original positive urine culture was collected) were recorded. Bacteremia was defined as any bacterial or fungal growth from a blood culture specimen, excluding coagulase-negative staphylococci and corynebacteria, as these could potentially have been skin contaminants. To explore the relationship between bacteriuria and bacteremia and to determine whether the bacteremia was potentially from a urinary source, we classified each episode of bacteremia as a definite match, a possible match, or discordant from the organisms isolated in the urine. A definite match meant that the urine and blood organisms were identical on a genus-and-species level, if species-level identification was available. A possible match meant that we had inadequate information to declare a definite match but that the organisms in the blood and urine might be the same. For example, if the urine grew gram-positive organisms, not further identified, and the blood grew Staphylococcus aureus, this would be a possible match. When the organisms clearly differed, for example, urine culture with a gram-negative organism and blood culture with S. aureus, we classified these cultures as discordant. If the blood cultures were done on the day before or the day of the urine culture, the bacteriuria was considered concurrent with the bacteremia. If the blood cultures were done from 1 to 30 days after the urine culture, the bacteriuria was considered prior to the bacteremia. We studied all-source bacteremia, not just urinary-source bacteremia, as the broad-spectrum antibiotics often used empirically to treat suspected CAUTI might have a beneficial effect in treating organisms beyond a urinary source.
Determining Antimicrobial Treatment
Data on all antimicrobials prescribed from 7 days before to 30 days after the original urine culture for all urine culture episodes were extracted from the pharmacy database of VA Decision Support System Clinical National Data Extracts.21 We also reviewed the electronic medical records to determine whether the antimicrobial agents given were intended to treat urinary organisms. We defined 2 variables for antimicrobial use: those given specifically for treatment of bacteriuria (defined as antimicrobials given for bacteriuria) and any antimicrobials given from 7 days before to 30 days after the urine culture (defined as any antimicrobials). The “any antimicrobials” category includes “antimicrobials given for bacteriuria.”
Univariate analysis was used to test for differences in characteristics associated with CAUTI and CAABU. A t test was performed for continuous variables, Fisher’s exact test was used to analyze categorical variables with 2 possible categories, and the χ2 test was used for categorical variables with more than 2 categories.
Multivariable analysis was used to explore Aim 2, the relationship between urinary symptoms (CAUTI vs CAABU) and the 2 outcomes of bacteremia from any source and mortality. Hierarchical logistic regression with nesting of urine culture episodes within patients was used to analyze cases of bacteremia from any source, whereas mortality analysis was based on individual patients rather than episodes of bacteriuria. Independent variables included the classification of catheter-associated bacteriuria as CAABU or CAUTI; demographic and clinical characteristics, including age, race, and Charlson comorbidity score; catheter type and duration; and the category of organisms in the urine. Urine microbiology results, including polymicrobial cultures, were placed into 3 mutually exclusive and exhaustive categories: (1) any gram-negative organisms present (with or without other species), (2) Candida alone present (no other organisms), and (3) all remaining urine culture results (eg, mixed gram-positive organisms). To explore our third aim, the effect of antimicrobials on the 2 outcomes of bacteremia from any source and mortality, we added a variable for antimicrobial use to the independent variables described above for Aim 2. To assess whether our classification of catheter-associated bacteriuria as CAABU or CAUTI might have resulted in a change in the significance of our findings, we also ran these models while excluding the CAABU-versus-CAUTI variable. All data analyses used SAS statistical software, version 9.2 (SAS Institute). A P value of .05 or less was considered statistically significant.
Numbers of Patients and Episodes of Bacteriuria
We captured 444 cases of catheter-associated bacteriuria in 308 unique patients. Table 1 describes the baseline characteristics of the 308 unique patients, based on the last urine culture sample collected from each patient. The mean age of patients was 72.3 years (standard deviation, 11.4), and the mean Charlson comorbidity score was 4.1 (standard deviation, 3.2). The study sample includes 253 (82.1%) patients from the medical wards and 55 (17.9%) from the extended care wards. Of the 308 patients included in our study, all with urinary catheters of some type, 45.5% of the patients in the extended care unit had indwelling catheters, whereas 60.1% in medical care units had indwelling catheters (, χ2). Among study patients, 277 of 308 (89.9%) received some antimicrobial agent during the period from 7 days before to 30 days after the urine culture. Overall mortality was 21.1% (61 of 308 patients) within 30 days. Among the 444 episodes of positive urine culture, an indwelling catheter was associated with 258 (58.1%) and an external catheter was associated with 186 (41.9%); 285 (64.2%) of these urine cultures had at least one gram-negative organism, 58 (13.1%) had Candida alone, and 101 (22.8%) had either exclusively gram-positive organisms or both gram-positive organisms and Candida. In order of frequency, the symptoms used to make a diagnosis of CAUTI rather than CAABU were fever, delirium, hematuria, dysuria, suprapubic tenderness, increased frequency, pelvic discomfort, flank pain, urgency, and rigors.
|Characteristic (N = 308)||Patients with CAUTI (n = 128)||Patients with CAABU (n = 180)||P value|
|≥65||88 (68.8)||121 (67.2)|
|<65||40 (31.3)||59 (32.8)|
|Male||123 (96.1)||175 (97.2)|
|Female||5 (3.9)||5 (2.8)|
|White||65 (50.8)||88 (48.9)|
|Black||49 (38.3)||72 (40.0)|
|Hispanic||8 (6.3)||10 (5.6)|
|Other/unknown||6 (4.7)||10 (5.6)|
|Charlson comorbidity score, mean (SD)||3.79 (3.01)||4.25 (3.35)||.22|
|Urinary catheter type||.30|
|Indwelling||78 (60.9)||99 (55)|
|External||50 (39.1)||81 (45)|
|Any antimicrobials received||125 (97.7)||155 (86.1)||<.001|
|Antimicrobials received for bacteriuria||117 (91.4)||90 (50.0)||<.001|
|Mortality within 30 days||21 (16.4)||44 (24.4)||.09|
Description of Relationship between Bacteriuria and Bacteremia from a Urinary Source
Among the 444 episodes of positive urine culture, 52 (11.7%) showed bacteremia from any source in the period from 1 day before to 30 days after the urine culture (after 36 episodes in which the organism in the blood culture was either coagulase-negative staphylococci or corynebacteria were excluded). Details of 9 (2% of 444) episodes of bacteriuria in which bacteriuria preceded the bacteremia by at least 1 day and in which the organisms in both blood and urine were either definite or possible matches appear in Table 2. These represent the cases of bacteremia for which treatment targeted at the organism in the urine might presumably have prevented subsequent bacteremia. However, 6 of the 9 cases had, upon further chart review, most likely a nonurinary source (Table 2), leaving 3 (0.7% of 444) cases of bacteremia from a urinary source, out of 444 episodes (0.7%), that could possibly have been prevented. One of these 3 cases was CAABU, while the other 2 were CAUTI.
|Episode||Organism(s) on |
|Days between |
|Suspected infectious |
source of bacteremia
|Other conditions||Antimicrobials |
|Outcome at |
|1||GPO||S. aureus||7||CAUTI||Balanitis: penile swab grew S. aureus||Hypopharyngeal cancer, diabetes||Yes||Alive|
|CAABU||PICC line (Proteus), decubitus ulcer||Colorectal cancer (resolved), dementia||No||Alive|
|3||E. coli, GPO||S. pneumoniae||6||CAUTI||Pneumonia||Hypertension, osteoporosis||Yes||Dead|
|CAABU||Cholecystitis, necrotizing pneumonia; PICC line (Candida)||Cirrhosis||No||Alive|
|5||GPO||S. aureus||8||CAABU||Probable endocarditis||Diabetes, renal cell carcinoma||No||Alive|
|7||GNR, GPO||Enterobacter||29||CAUTI||Urinary||Cirrhosis, diabetes||No||Alive|
|8||Candida, GPO||S. aureus||16||CAUTI||Endocarditis||Cirrhosis, diabetes||No||Alive|
|9||E. coli||E. coli||1||CAUTI||Urinary||Chemotherapy and steroids for lymphoma||Yes||Alive|
Association of Urinary Symptoms (CAUTI or CAABU) with Bacteremia or Mortality
Of 52 episodes of bacteremia from any source within 30 days, 32 (61.5%) occurred with episodes of CAUTI, and 20 (38.5%) occurred with episodes of CAABU. CAUTI, rather than CAABU, was significantly associated with bacteremia from any source, while having an external catheter rather than an indwelling catheter was protective (see Table 3). The variables significantly associated with mortality were a higher Charlson comorbidity score and having a pure growth of Candida versus a gram-negative organism in the urine culture (see Table 3). CAUTI versus CAABU was not associated with mortality.
|Independent variables||Odds ratio||95% CI||Odds ratio||95% CI|
|CAUTI vs CAABU||2.8||1.24–6.22||0.8||0.40–1.60|
|Antimicrobials given for bacteriuria||0.7||0.32–1.71||0.9||0.43–1.69|
|Charlson comorbidity score||0.9||0.66–1.26||1.1||1.01–1.21|
|Urinary catheter type: external vs indwelling||0.5||0.22–0.99||1.5||0.84–2.84|
|Organisms on urine culture|
|Candida in pure culture vs gram-negative rod||0.9||0.31–2.49||3.6||1.57–8.26|
|All remaining organisms vs gram-negative rod||1.0||0.43–2.28||1.2||0.58–2.51|
|Catheter duration: <2 weeks vs >2 weeks or prior to emergency room||1.3||0.63–2.58||0.8||0.43–1.47|
|Black vs white||0.9||0.42–1.81||1.1||0.57–1.94|
|Hispanic vs white||0.5||0.09–2.50||0.8||0.22–3.23|
|Other/unknown vs white||1.4||0.37–5.52||1.6||0.46–5.33|
Effect of Antimicrobial Treatment for Bacteriuria on Bacteremia and Mortality
Administration of antimicrobials for bacteriuria did not affect the risk of bacteremia from any source or mortality within 30 days. We reran the analysis after removing the CAABU-versus-CAUTI variable and again did not find an association between the use of antimicrobials for bacteriuria and bacteremia from any source or mortality. Since all 52 bacteremia episodes were associated with prescription of at least one antimicrobial agent within the period from 7 days before to 30 days after the original positive urine culture, we were unable to analyze whether withholding antimicrobials entirely would have affected the occurrence of bacteremia from any source. When we repeated the multivariate analysis for mortality using the variable of any antimicrobial use from 7 days before to 30 days after the urine culture, there still was no difference in mortality at 30 days between the 2 groups ().
We studied the difference in outcomes between CAUTI and CAABU, as defined by IDSA guidelines. CAUTI, rather than CAABU, was significantly associated with bacteremia from any source but not with mortality. Administration of antimicrobials for bacteriuria did not affect the risk of bacteremia from any source or mortality within 30 days within this group of patients with substantial comorbidities and high underlying antimicrobial use. The lack of effect of antimicrobials was not affected by whether we included presence or absence of urinary symptoms as a variable.
The purpose of treating CAUTI is 2-fold: to relieve symptoms7 and to prevent bacteremia from a urinary source and, potentially, sepsis and death.13,22 We found that only 3 of the 52 bacteremias from any source (or 3 of 444 [0.68%] catheter-associated bacteriuria episodes) were actually from a urinary source and potentially preventable by treating the urine organism. Our findings are similar to those of Tambyah and Maki,3 for whom 1 (0.43%) of 235 patients with bacteriuria developed bacteremia from a urinary source. Also, as in the Tambyah and Maki study, 1 of these 3 bacteremias was asymptomatic, with CAABU rather than CAUTI. Our findings raise the question of whether symptom-based classification of cases as CAUTI rather than CAABU is relevant to the use of antimicrobials to prevent bacteremia from a urinary source. Others have suggested tailoring antibiotic therapy to certain patient subgroups that may be at increased risk of bacteremia.13-15 Our finding that having CAUTI rather than CAABU was significantly associated with bacteremia from any source was not surprising, because one criterion we used for classifying cases of catheter-associated bacteriuria as CAUTI was the presence of fever in the absence of another clear etiology. We found a potentially protective effect of the use of an external catheter rather than an indwelling catheter in terms of bacteremia from any source. This finding should be interpreted with awareness that patients’ treatment was often switched from an indwelling catheter to an external catheter and vice versa during the course of follow-up. A plausible biologic explanation is that the external catheters are less likely to cause mucosal trauma than indwelling catheters, thus decreasing the risk of bacteremia.
Our study population had a high mortality rate (21.1%), comparable to the mortality seen among catheterized patients with bacteriuria in past studies (18%–43%).4-6 Our study population had a high comorbidity score, with an average of 4.1, and a higher Charlson comorbidity score was associated with mortality. Given that patients were admitted with a variety of conditions and that they were not in an intensive care unit at the time of the urine culture, there was no overall score for severity of acute illness that we could have used in our analysis, and we did not examine the reported cause of death. Having a pure growth of Candida versus a gram-negative organism in the urine culture was also associated with mortality within 30 days. An association between candiduria and crude mortality has been found in multiple other studies of hospitalized patients and is likely a reflection of heavy prior antimicrobial use in patients with substantial underlying comorbidities.23-27
One limitation of our retrospective study design is that classification of urine cultures as CAUTI or CAABU was based on chart review. Providers who added antibiotics to patients with bacteriuria usually provided justification in the notes, documenting whether the antimicrobial agents were added for a urinary or a nonurinary infectious source, and we used this to determine the indication for antimicrobial use. Although it would have been ideal to further clarify antimicrobials as appropriate or inappropriate on the basis of matches to the susceptibility of organism(s) in the urine, this was impractical, given that the catheter-associated bacteriuria was often polymicrobial and that patients’ treatment was often switched from one antimicrobial to another. It was disheartening to note that 50% of the patients with CAABU received antimicrobials targeting their urinary tract organisms. Our study results cannot be generalized to all patients with urinary catheters, as most of our patients were men, and condom catheter use may have been more common in this facility than in other settings. The strengths of the study include the emphasis on distinguishing CAUTI from CAABU and the relatively large sample of episodes of bacteriuria. Given the low number of cases of bacteremia of urinary origin, a multicenter study would be necessary to focus specifically on that outcome. We also looked at the greater context of all antimicrobial agents used in our study population, taking into account that patients who develop catheter-associated bacteriuria are likely to receive antimicrobial agents regardless of whether they receive specific treatment for urinary flora.
Our findings document that bacteremia from a urinary source is infrequent in patients with catheter-acquired bacteriuria, and there was no evidence of an association of all-cause mortality with symptomatic versus asymptomatic bacteriuria in this population. Furthermore, antibiotic treatment of bacteriuria, symptomatic or asymptomatic, was not associated with decreased mortality.
We thank Armandina Garza, Ashley Gendrett, and Lisa Hughes for assisting with data collection and verification and Annette Walder for programming assistance.
Financial support. This work was supported by grants from the Department of Veterans Affairs (VA Rehabilitation Research and Development Career Development Award B4623 and VA Health Services Research and Development Investigator-Initiated Research award 09–104) and the National Institutes of Health (NIH DK092293) to B.W.T. This manuscript is the result of work supported with resources and use of facilities at the Houston VA Health Services Research and Development Center of Excellence (HFP90-020) at the Michael E. DeBakey VA Medical Center, Houston, Texas.
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article. All authors submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and the conflicts that the editors consider relevant to this article are disclosed here.
- 1. Klevens RM, Edwards JR, Richards CL Jr, et al. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep 2007;122:160–166.
- 2. Siddiq DM, Darouiche RO. New strategies to prevent catheter-associated urinary tract infections. Nat Rev Urol 2012;9:305–314.
- 3. Tambyah PA, Maki DG. Catheter-associated urinary tract infection is rarely symptomatic: a prospective study of 1497 catheterized patients. Arch Intern Med 2000;160:678–682.
- 4. Platt R, Polk BF, Murdock B, Rosner B. Mortality associated with nosocomial urinary-tract infection. N Engl J Med 1982;307:637–642.
- 5. Clec’h C, Schwebel C, Français A, et al. Does catheter-associated urinary tract infection increase mortality in critically ill patients? Infect Control Hosp Epidemiol 2007;28:1367–1373.
- 6. Laupland KB, Bagshaw SM, Gregson DB, Kirkpatrick AW, Ross T, Church DL. Intensive care unit-acquired urinary tract infections in a regional critical care system. Crit Care 2005;9:R60–R65.
- 7. Hooton TM, Bradley SF, Cardenas DD, et al. Diagnosis, prevention, and treatment of catheter-associated urinary tract infection in adults: 2009 international clinical practice guidelines from the Infectious Diseases Society of America. Clin Infect Dis 2010;50:625–663.
- 8. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309–332.
- 9. Nicolle LE, Bradley S, Colgan R, Rice JC, Schaeffer A, Hooton TM. Infectious Diseases Society of America guidelines for the diagnosis and treatment of asymptomatic bacteriuria in adults. Clin Infect Dis 2005;40:643–654.
- 10. Warren JW, Anthony WC, Hoopes JM, Muncie HL Jr. Cephalexin for susceptible bacteriuria in afebrile, long-term catheterized patients. JAMA 1982;248:454–458.
- 11. Tenke P, Kovacs B, Bjerklund Johansen TE, Matsumoto T, Tambyah PA, Naber KG. European and Asian guidelines on management and prevention of catheter-associated urinary tract infections. Int J Antimicrob Agents 2008;31(suppl 1):S68–S78.
- 12. Leone M, Perrin AS, Granier I, et al. A randomized trial of catheter change and short course of antibiotics for asymptomatic bacteriuria in catheterized ICU patients. Intensive Care Med 2007;33:726–729.
- 13. Bahagon Y, Raveh D, Schlesinger Y, Rudensky B, Yinnon AM. Prevalence and predictive features of bacteremic urinary tract infection in emergency department patients. Eur J Clin Microbiol Infect Dis 2007;26:349–352.
- 14. Saint S, Kaufman SR, Rogers MA, Baker PD, Boyko EJ, Lipsky BA. Risk factors for nosocomial urinary tract-related bacteremia: a case-control study. Am J Infect Control 2006;34:401–407.
- 15. Greene MT, Chang R, Kuhn L, et al. Predictors of hospital-acquired urinary tract-related bloodstream infection. Infect Control Hosp Epidemiol 2012;33:1001–1007.
- 16. Burns AC, Petersen NJ, Garza A, et al. Accuracy of a urinary catheter surveillance protocol. Am J Infect Control 2012;40:55–58.
- 17. Trautner BW, Kelly PA, Petersen N, et al. A hospital-site controlled intervention using audit and feedback to implement guidelines concerning inappropriate treatment of catheter-associated asymptomatic bacteriuria. Implement Sci 2011;6:41.
- 18. Trautner BW, Bhimani RD, Bush AL, et al. Development and validation of an algorithm to recalibrate mental models and reduce diagnostic errors associated with catheter-associated bacteriuria. BMC Med Inform Decis Making 2013;13:48.
- 19. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med 2005;37:360–363.
- 20. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–174.
- 21. Smith MW, Joseph GJ. Pharmacy data in the VA health care system. Med Care Res Rev 2003;60(3 suppl):92S–123S.
- 22. Sedor J, Mulholland SG. Hospital-acquired urinary tract infections associated with the indwelling catheter. Urol Clin North Am 1999;26:821–828.
- 23. Kauffman CA, Vazquez JA, Sobel JD, et al. Prospective multicenter surveillance study of funguria in hospitalized patients. Clin Infect Dis 2000;30:14–18.
- 24. Paul N, Mathai E, Abraham OC, Michael JS, Mathai D. Factors associated with candiduria and related mortality. J Infect 2007;55:450–455.
- 25. Álvarez-Lerma F, Nolla-Salas J, León C, et al. Candiduria in critically ill patients admitted to intensive care medical units. Intensive Care Med 2003;29:1069–1076.
- 26. Magill SS, Swoboda SM, Johnson EA, et al. The association between anatomic site of Candida colonization, invasive candidiasis, and mortality in critically ill surgical patients. Diagn Microbiol Infect Dis 2006;55:293–301.
- 27. Toya SP, Schraufnagel DE, Tzelepis GE. Candiduria in intensive care units: association with heavy colonization and candidaemia. J Hosp Infect 2007;66:201–206.