THE JOURNAL OF INFECTIOUS DISEASES • VOL. 155, NO.5. MAY 1987 Diagnosis of Bacteriuria in Men: Specimen Collection and Culture Interpretation

Bacteriologic criteria for diagnosing urinary tract infections in men have not been well defined. For determination of the bacterial colony count in voided specimens that most accurately reflects bladder bacteriuria, culture results were compared for voided urine and bladder urine (obtained by suprapubic aspiraton and urethral catheterization) from men with various genitourinary problems. Bladder bacteriuria was found in 36 (47.3%) of 76 sets of specimens from 66 individuals. Culture results of bladder specimens showed excellent agreement with those of clean-catch midstream-void and uncleansed first-void specimens (weighted K = 0.924 and 0.906, respectively). The criterion for clean-catch midstream-void specimens that best differentiated sterile from infected bladder urine was growth of ~103 cfu of one predominant species/ml; this definition had a sensitivity of 0.97 and a specificity of 0.97. Uncleansed first-void specimens were equally sensitive (O.97) but less specific (O.91-0.92)in detection of bacteriuria. Pyuria (~1O leukocytes/mrn") and irritative genitourinary symptoms showed modest correlations with bladder bacteriuria. Nearly all studies examining bacteriologic criteria for diagnosing urinary tract infections have been per-formed in predominantly or exclusivelyfemale popu-lations, even '.iough the prevalence of bacteriuria in elderly men approaches that in women (i,e., 20070-30010) [1,2]. Standards derived from these studies in women have been empirically applied to men, de-spite important differences in their genitourinary anatomy [2-4]. Investigations of the epidemiology, pathogenesis, and treatment of urinary tract infec-tions in men require well-validated methods for de-tection of bacteriuria [3, 5]. Therefore we sought to Received for publication 2 June 1986, and in revised for