Natural history and epidemiology of benign prostatic hyperplasia: Relationship among urologic measures

Numerically weak correlations can stem from research studies for numerous reasons, some of which have little to do with lack of true physiologic relationships. Previous investigators have reported relatively weak correlations among urologic measures, mostly based on patients referred to urologic clinics. Such samples of patients may be prone to self-selection or referral bias. Men seeking medical care for urinary symptoms are more likely to have more severe symptoms, low urinary flow rates, and prostatic enlargement, and hence clinic-based samples may reflect a narrow spectrum in urologic measures, resulting in attenuated correlation coefficients. Measurement error in the technique or equipment, lack of specificity, and within-patient variability can also attenuate correlations. Thus, the characteristics of the sample upon which correlations are calculated must be considered in the interpretation of the magnitude of relationships among measurements. Possibly due to the broader spectrum, community-based studies have recently found modest but somewhat stronger relationships among urologic measures than previously documented. Such correlations are comparable to those found in other disease areas. In addition, such studies have documented that symptoms, urinary flow rate, and prostatic enlargement are predictive of long-term complications, such as acute urinary retention, with predictive relationships comparable to those found in other disease areas. Thus, the relationships among and predictive ability of urologic measures may be stronger than currently appreciated. (C) 1998, Elsevier Science Inc. All rights reserved.