The principle finding of this study is that in adult patients with acute GBS UTI, while mean s-QBC values are higher than in patients with ABU, these measures are not useful in distinguishing between these conditions. These findings are noteworthy given the widespread application of s-QBC values for identifying acute UTI in patients who are culture-positive for various organisms, including GBS. The finding that s-QBC is not a significant prognostic indicator for acute GBS UTI has practical implications in terms of disregarding low s-QBC values in GBS-positive patients. Low-count s-QBC is sometimes associated with acute infection
, and the data reported here suggest that interpreting low-count s-QBC in GBS-positive patients as merely transient flora, or contaminants is inappropriate. Thus, nominal s-QBC cut-off values commonly used such as ≥105 CFU/ml and ≥104 CFU/ml should be considered as insensitive for the diagnosis of acute GBS UTI. The data presented here suggest that a cut-off of ≥103 CFU/ml might be appropriate (the limit of detection for most laboratories). This is consistent with recent ideas highlighting the value of lower s-QBC cut-offs in the setting of adult patients infected with other uropathogens
Semi-quantitative urinary cultures have been reported in several studies in which over half of symptomatic patients exhibited s-QBC values lower than the traditional cut-off of 105 CFU/ml often used for defining significant bacteruria
[35, 36]. Others have reported that a ≥102 CFU/ml s-QBC criterion may be superior to a ≥105 threshold
[37, 38]. In a report by Thomsen et al. that investigated penicillin treatment of patients with GBS s-QBC of any count, there was a reduction in poor pregnancy outcomes compared to a placebo control, also suggesting that s-QBC lower than the traditional cut-off of 105 CFU/ml are clinically important
. These findings show that s-QBC values as low as 102 CFU/ml are especially significant in pregnant women. In the broader context, women with low-grade bacteruria due to other uropathogens have been shown to respond to therapy
 supporting the notion that low-grade s-QBC values are important for uropathogens in general.
It is important to highlight the limitations of the current study, and the possible implications of lowering the s-QBC cut-off value for interpreting GBS culture-positive patients in the diagnostic setting of UTI. Contamination of specimens in the current work could not be suitably addressed due to the retrospective nature of the study and the unfeasibility of repeat urine cultures and catheterized samples for all patients in this setting. Overall, the number of catheterized samples was low (~10%), and related to both acute UTI and ABU patients. We defined ABU as the isolation of urinary GBS in pure culture from a single sample, which is an approach consistent with prior studies
[14, 41]. Others have limited the definition of ABU to cases in which two- or three-consecutive samples from an individual are culture-positive for the same or similar organism
[32, 42, 43]. Our goal was to incorporate the highest number of cases possible into our study, and its retrospective design ruled out repeat sampling. Thus, some of the asymptomatic women with urinary GBS in this study would represent cases of sample contamination, especially for low count samples, whilst others may reflect low-grade bacteruria. Despite this limitation, data related to symptomatic patients with low-grade s-QBC, and thus, the key finding of the study, remain valid. Future studies of subjects with repeat samples incorporating a higher proportion of catheterized patients would be useful to address this.
Contamination and recovery of multiple organisms occurs in approximately half of all urine specimens that are positive for GBS
, and we excluded all such cases in the current study. In contaminated specimens, other organisms may overwhelm low numbers of GBS causing a large number of UTIs due to this organism to be missed. There were also many patients with high GBS s-QBC but much lower numbers of other organisms in the current study (also excluded), which probably indicates a real infection but is discounted in the diagnostic laboratory as a contaminated specimen due to reporting. Overall, this means that many true GBS UTI cases were probably missed in this study (only single-organism cultures were used per case definitions). In the broader context, such cases continue to be missed in microbiology laboratories due to routine reporting practices in the diagnostic setting. Finally, a high prevalence of low-grade GBS bacteruria
 also means that revising down the s-QBC cut-off for GBS UTI would increase the rate of false positives. Despite these challenges, the data reported here show that a ≥105 CFU/ml threshold is insensitive as a marker for acute GBS UTI, which provides important new practical information for the diagnostic laboratory.
This study also suggests that age is a useful prognostic indicator for predicting acute GBS UTI, perhaps equally so as compared to s-QBC. Despite this, age still only identified 3/147 acute GBS UTI patients in our patient cohort based on a logistic regression model. The age measure was confounded in our study because many of the urine specimens obtained from asymptomatic women (~25-30%) were primarily related to perinatal screening of pregnant women. Nonetheless, patients over the age of 40 years were shown to be at significantly higher risk for acute GBS UTI compared to younger patients, suggesting age may be useful as a surrogate risk marker. This finding, against the background of high false positive rates for acute GBS UTI, highlights the need for careful interpretation of GBS-positive urine cultures. Using age as a prognostic tool could help to identify high-risk patients and partly address the difficulties in interpreting s-QBC values for GBS. Gender was not a prognostic factor for discriminating infection category, which may be more relevant to older persons and nursing home residents as discussed elsewhere
GBS infection in adults is often associated with diabetes, which is considered a risk factor for disease due to this organism
. The prevalence of ABU in diabetic individuals is higher compared to individuals without diabetes
, and although diabetic patients are more likely to have acute UTI due to any uropathogen, diabetes was not identified as a risk factor for GBS UTI in a prior study
. This suggests that the association between diabetes and symptomatic UTI may not be as relevant for GBS. The diabetic status of all study subjects in the current work was not included in the medical record review. Interestingly, however, the proportion of individuals in the current study with ABU who exhibited a positive test result for elevated urinary glucose was more than double that of patients with acute GBS UTI (22.4% vs. 10.1% respectively, a significant difference according to Pearson Chi-Square analysis) hinting at a possible connection. The role of diabetes in GBS UTI and possible bias in the current dataset due to differential pregnancy rates in different patient groups will require further investigation. The association between bacteruria and diabetes in relation to microbial sugar metabolism also raises the possibility of growth of uropathogenic organisms in urine
. Growth fitness in urine has been shown for E. coli, but steps in the pathogenesis of GBS UTI have only recently been described
[49–51], and a GBS isolate cultured from a patient with acute cystitis was unable to grow in urine
. Proteinuria in ABU patients in this study also implies that diabetes may be associated with GBS ABU. Comparing the proteinuria and hematuria data in ABU patients irrespective of esterase and pyuria UA findings (possible bias per the exclusion criteria) would also be of interest for further study.