New Participant Stratification and Combination of Urinary Biomarkers and Confounders Could Improve Diagnostic Accuracy for Overactive Bladder - Beyond the Abstract

Overactive bladder (OAB) is defined by urinary symptoms of urgency, increased frequency, nocturia, with or without urge incontinence, in the absence of proven infection or other obvious pathology. Prevalence is thought to be around one in five adults. Symptoms themselves are highly impactful on patients, their family members/carers and healthcare systems.

 
A number of observations suggest the need to better distinguish OAB from diseases with overlapping urinary symptoms. First, many clinicians will diagnose OAB based on the exclusion of these other diseases. Given flaws in the methods for a positive diagnosis of these other diseases (e.g. those used to confirm urinary tract infection), diagnosis by exclusion is a flawed approach. Second, while urodynamically-derived cystometric parameters provide some clinical value, detrusor overactivity itself does not indicate OAB. Furthermore, urodynamic assessment is not indicated in a proportion of patients (the frail elderly) and is an expensive procedure, given the requirement of a highly-skilled clinician. Our study, therefore, aimed to determine whether chemicals within urine could be used to predict OAB; a question that is necessary to answer if we are to develop accurate, non-invasive methods for diagnosis of OAB.

 

Many previous biomarker studies have compared participants at either end of the OAB symptom severity spectrum: asymptomatic ‘healthy’ controls vs. OAB patients with advanced symptoms (voiding every hour; waking several times during the night to void; often with urge incontinence). When we designed a study to look for urinary biomarkers of OAB, we chose not to compare two ends of the OAB spectrum because we thought that we’d only find biomarkers typical of the bladder remodeling that comes with a very unhealthy bladder. Therefore, we chose to recruit consented individuals (n=113) with a range of symptoms at the lower end of OAB symptom severity. Symptoms were recorded using the ICIQ-OAB questionnaire. Following screening to identify participants with infection, we processed the urine of 95 participants to analyze the concentrations of putative biomarkers (ATP, ACh, nitrite, MCP-1 and IL-5) and participants’ confounders (age and gender).

 

Participants occurred in two groups, based on the occurrence and severity of OAB symptoms: Group 1 reported virtually no urinary symptoms; Group 2 had mild to moderate symptoms of OAB. The separation of participants into groups was not arbitrary nor subjective; rather, we used a statistical technique called cluster analysis to detect naturally-occurring clusters. The presence and extent of urinary urgency was the biggest influence on this grouping.

 

In order to assess the individual and combinational abilities of candidate OAB urinary biomarkers plus participants’ confounders in predicting these two groups, we used binary logistic regression followed by receiver operating characteristic curve analysis. Six combination models were shown to have clinically acceptable diagnostic powers (AUC >0.7) and each had a higher negative predictive power (NPV) and positive predictive power (PPV) than detrusor overactivity. Amongst these combination models, a combination of age, gender, ATP and IL-5 was shown to have the highest PPV and NPV values and had a similar diagnostic performance as urine-based tests for other diseases such as bladder carcinoma¹ or prostate cancer.²

 

Using an approach to stratify participants based on OAB symptoms and characterize stratified groups based on urinary biomarkers, our study identified the means to both predict and exclude OAB. When tested in large, longitudinal studies, this approach and its findings offer the potential for the development of more accurate and non-invasive diagnostic tools.