We conducted a prospective cohort study of patients with enterococcal bacteriuria admitted to Barnes–Jewish Hospital (BJH) between 1 July 2011 and 31 December 2011. BJH, a 1250-bed teaching hospital, is the largest hospital in Missouri, and has a referral base that includes the St Louis metropolitan area, eastern Missouri and western Illinois. It houses all medical specialties. BJH is affiliated with Washington University School of Medicine.

Patients with enterococcal bacteriuria were identified by means of a daily query of the hospital medical informatics system by one of the authors (H.K.). We included all adult inpatients with enterococcal bacteriuria (aged ≥17 years) and excluded outpatients or patients discharged home directly from the emergency department and those with concurrent non-enterococcal bacteraemia within 3 days of enterococcal bacteriuria. Polymicrobial bacteriurias were not excluded; they were defined as detection of two or more microbial species in the same urine culture.

Upon inclusion, one of the authors (H.K.) reviewed the patients' medical records to identify signs and symptoms associated with the bacteriuria. Demographic and clinical data were also collected plus pertinent laboratory and radiological findings. Both antibiotic management (including antibiotic prescriptions upon discharge) and urinary catheter management were recorded. Enrolled patients' medical records were reviewed daily during admission and the following outcomes were evaluated: (i) symptom resolution within 3 days of the start of treatment (or at the time of discharge if discharged within 3 days of the positive culture); (ii) secondary bloodstream infection; (iii) length of hospital stay; (iv) transfer to the intensive care unit (ICU); and (v) crude in-hospital mortality. Post-discharge outcomes of interest were readmission to the hospital within 30 days with recurrent bacteriuria and/or bacteraemia.

Routine microbiological work-up on the specimens was performed by the BJH Clinical Microbiology Laboratory, including speciation (into E. faecalis, E. faecium, or other enterococcal species) and testing for antimicrobial susceptibilities. The latter was done with disc diffusion methodology, using cut-offs proposed by the Clinical and Laboratory Standards Institute (http://www.clsi.org). The cut-off for significant bacteriuria employed in our hospital microbiology laboratory was 5 × 10⁴ colony-forming units (CFU)/mL in non-catheterized and 5 × 10³ CFU in catheterized patients.

Symptomatic UTI was defined as the presence of bacteriuria and one or more bladder symptoms (dysuria, hesitancy, frequency, suprapubic pain), flank pain and/or fever. Cystitis was defined as the presence of bladder symptoms, including dysuria, hesitancy, frequency, and suprapubic pain. Pyelonephritis was defined as the presence of fever and/or flank pain in a bacteriuric patient. Asymptomatic bacteriuria (ASB) denoted the absence of any of the above-mentioned symptoms.¹¹ Patients unable to report symptoms were considered unclassifiable. Bacteriuria was considered to be catheter-associated if a urinary catheter was present within 48 h before the positive urine culture. Pyuria was defined as >10 white blood cells per high-power field in urine microscopy.

In 2010, 28% of enterococci recovered from clinical urine cultures at BJH were vancomycin resistant. We hypothesized that a 15% difference in successful clinical treatment (i.e. resolution of symptoms at day 3 or at discharge) between VRE and VSE would be clinically significant. Assuming that symptom resolution is achieved in 80% of VRE bacteriurias vs 95% of VSE bacteriurias it was estimated that 55 vs 141 patients (or total, 196) would be required to make a statistically significant statement (given alpha = 0.05 and power = 80%).

We used the statistical package SPSS (SPSS Inc., version 18, Chicago, IL, USA) to perform analyses. Univariate analyses included chi-square test or Fisher's exact test for categorical variables, as appropriate, and Student's t-test or Mann–Whitney U-test for continuous variables, as appropriate. P < 0.05 was considered statistically significant. Variables with P ≤ 0.1 were included in a multivariate model. Outcomes were compared between VRE and VSE bacteriurias by means of the chi-square test. The project was approved by the Human Research Protection Office at Washington University (which waived informed consent due to the observational nature of the study) on 25 May 2011. The submission number was 201104218.

In all, 254 patients with enterococcal bacteriuria were enrolled, of whom 160 (63%) were female and 171 (67%) white. The median age was 65 years (range: 17–96 years). The body mass index (BMI) was 26.3 kg/m² (range: 15.7–79.6). The clinical manifestations ranged from ASB (119; 47%), to pyelonephritis (64; 25%), cystitis (20; 8%), and unclassified bacteriuria (51; 20%); 116 (46%) had hospital-acquired bacteriurias, and 145 (57%) were catheter-associated bacteriurias. Table I summarizes the baseline demographic characteristics and comorbidities of these patients, depending on vancomycin susceptibility. Seventy-four (29%) urine isolates were vancomycin resistant (VRE). Compared to those with VSE bacteriuria, patients with VRE bacteriuria were more likely to be non-white (P = 0.02), to have renal insufficiency (P = 0.001), and to be receiving immunosuppressive treatment (corticosteroids, other immunosuppressants, chemotherapy). In multivariate analysis, non-white race [odds ratio (OR): 1.96; 95% confidence interval (CI): 1.05–3.70] and renal insufficiency (OR: 2.37; 95% CI: 1.27–4.41) were independent predictors of vancomycin resistance. Otherwise the two groups were similar in their demographics and comorbidities. A total of 177 (70%) patients received pathogen-directed antibiotic treatment.

Out of 254 isolates, 243 were available for speciation (96%). The majority of these were E. faecalis (176; 69%), followed by E. faecium (62; 24%), and other Enterococci spp. (5; 2%). E. faecalis were resistant to vancomycin in 32% of cases, whereas 79% of E. faecium were VRE. There were 61 (24%) polymicrobial bacteriurias. Polymicrobial bacteriurias were associated with functional or anatomical urinary tract abnormality (P = 0.01) and with pyuria (P = 0.03), but did not result in worse outcomes than monomicrobial infections (data not shown).

A significant number of patients had ASB (119; 47%), of which 70 (59%) received antibiotic treatment. Of the 70 who received antibiotic treatment, 24 (34%) had a concomitant diagnosis of infection at another site that could have led to antibiotic usage. Based on the medical record, we could not determine whether antibiotics were given to treat the bacteriuria or for the alternative indication in these patients; 46 (39%) of 119 ASB patients received antibiotics exclusively for ASB. The presence of pyuria in patients with ASB did not affect the likelihood of antibiotic administration [35/54 (65%) with pyuria vs 33/65 (51%) without pyuria; P = 0.1]. Among antibiotics used for ASB, vancomycin (14/70; 20%), ampicillin, and linezolid (each 13/70; 19%) were the most common. Vancomycin resistance was not associated with increased antibiotic use in ASB [15/28 (54%) in VRE vs 55/91 (60%) in VSE; P = 0.5].

Of the 254 urine isolates tested, 74 (29%) were VRE. Overall, symptoms were not significantly more likely to be reported in VRE vs VSE infections [46/137 (34%) vs 28/121 (24%); P = 0.06]. The two groups were also similar with regard to the subsets of cystitis, pyelonephritis, and ASB (Table II).

Out of 254 patients, 177 (70%) received pathogen-directed antibiotics. Linezolid was more likely to be used as the first pathogen-directed antibiotic in patients with VRE bacteriuria [25 (51%) vs 9 (7%) in VSE bacteriuria patients; P < 0.001]. Also, doxycycline [7 (14%) vs 5 (4%); P = 0.02] and daptomycin [4 (8%) vs 0 (0%); P = 0.005] were more commonly used in patients with VRE infection. Twenty-four (32%) of VRE isolates were ampicillin susceptible; 7 (29%) were treated with linezolid. There were no significant differences in other antibiotics used. As expected, vancomycin was not used as pathogen-directed antibiotic for VRE. In addition to antibiotic management, catheter management was examined across the two groups: in the VRE group, 23 (55%) patients had catheters removed compared to 47 (45%) in the VSE group (P = 0.3).

Vancomycin susceptibility did not significantly alter outcomes such as transfer to ICU [10 (14%) VRE vs 17 (9%) VSE; P = 0.3], length of hospital stay [6.8 days (0.1–150.8) vs 5 days (0.2–78.2); P = 0.08], and mortality [10 (14%) vs 13 (7%); P = 0.1] (Table II). Outcomes were similarly unaffected in the symptomatic subset of patients (data not shown). Bacteraemia was present in 8 (3%) of the bacteriuria cases. Thirty-day post-discharge outcomes were similar independent of vancomycin resistance.