Tracking of Longitudinal Changes in Measures of Benign Prostatic Hyperplasia in a Population Based Cohort

Abstract

We characterized how longitudinal changes in PSA, prostate size, maximum urinary flow rates and lower urinary tract symptoms track together over time. In 1990, 2,115 white men, randomly selected from the Olmsted County, Minnesota population, completed validated questionnaires during a home visit by a study assistant. A 25% random subsample underwent clinical evaluation including transrectal ultrasonography, serum PSA and assessment of maximum urinary flow rates. Examinations and questionnaires were repeated biennially through 2002. Longitudinal changes in these measurements were estimated with 2-stage models. Annualized changes were correlated and also dichotomized at various percentiles and examined in age adjusted logistic regression models predicting symptom increases in the upper 80th percentile. Correlations between changes in prostate volume, PSA levels, maximum flow rates and urinary symptoms were modest (age adjusted Spearman correlation coefficients: volume and symptoms 0.08, p = 0.06; PSA and symptoms 0.06, p = 0.20; maximum flow rate and symptoms -0.08, p = 0.05). However, PSA and prostate volume annual increases in the upper 80th percentile were each associated with an approximately 2-fold increased risk of symptom changes in the upper 80th percentile. As PSA and prostate volume changes increased from the 50th to the 90th percentiles, the odds of having symptoms in the upper 80th percentile also increased. While overall correlations among changes in each of these variables were modest, men with more rapid annual increases in PSA levels and prostate volumes were also likely to have more rapid increases in urinary symptoms.