SYMPTOM?BASED CLUSTERING OF MALE PARTICIPANTS IN THE SYMPTOMS OF LOWER URINARY TRACT DYSFUNCTION RESEARCH NETWORK (LURN) OBSERVATIONAL COHORT STUDY
INTRODUCTION
The identification of subtypes of patients with lower urinary tract symptoms (LUTS) is critical to the understanding of pathobiology and effective clinical management. Conventional classification of LUTS patients based solely on the predominant individual symptoms may be unsatisfactory due to the multitude of potential causes and processes. We propose a novel method of unsupervised classification based on detailed multi-symptom information as an alternative to traditional diagnostic categories.
METHODS
We analyzed baseline data from 503 care-seeking male participants in the LURN Observational Cohort study. Symptoms were measured using the LUTS Tool and AUA Symptom Index (52 questions total). A probability-based consensus clustering algorithm identified patient subtypes with distinct symptom signatures.
RESULTS
Four distinct clusters were identified (Figure). Patients in cluster 1 (n=166) had predominant symptoms of frequency, hesitancy, straining, weak stream, intermittency and incomplete bladder emptying, consistent with &[Prime]classic&[Prime] bladder outlet obstruction. Patients in cluster 2 (n=93) endorsed predominant post-micturition symptoms (e.g., post-voiding dribbling), with some stream alterations. Patients in cluster 3 (n=114) endorsed predominant frequency and some urgency without urinary incontinence; this distinguished cluster 3 from cluster 4 (n=130) wherein patients reported similar frequency and urgency, but with more severe urinary incontinence. All but two symptom questions were statistically different between at least two clusters; each symptom-based cluster contained at least one patient from each conventional symptom subtype (continent, post-void dribble/post-void urinary incontinence, and incontinent).
CONCLUSION
We identified four data-derived male LUTS subtypes that differ from conventional subtypes among men seeking LUTS care. Comment: Validation of these subtypes will determine if clusters guide clinical management. Further subtype refinement will incorporate clinical assessments and non-urinary patient-reported outcomes. Anatomical, physiological and psychosocial testing may lead to better understanding of LUTS pathophysiology and treatment.
Funding: Funding provided by grants from NIH/NIDDK