The clinical problems and results of urinalyses of 500 dogs were reviewed and summarized to compare the sensitivities for detection of abnormalities indicative of urinary system disease among qualitative (sulfosalicylic acid [SSA]), quantitative (Coomassie brilliant blue [CBB]), and indexed (urinary protein/creatinine ratio [U(P/C)]) determinations of urinary protein loss vs microscopic examination of urine sediment. False-negative rates for the detection of microscopically abnormal urine specimens were 5.4% for SSA greater than or equal to 1+, 8.5% for CBB greater than or equal to 1.0 mg/ml, 9.7% for U(P/C) greater than or equal to 1.0, and 7.7% for CBB + U(P/C). A discriminatory U(P/C) value of 2.0 would have excluded dogs with clinically relevant proteinuria in the lower ranges. Proteinuria was not detected in 4.4% (22/500) of the specimens in which important numbers of leukocytes or bacteria were observed. False-negative rates for combined interpretation of quantitative protein concentration and U(P/C) were not significantly different (P greater than 0.10) from SSA alone. Degrees of azotemia were higher (high serum creatinine concentration, P greater than 0.10 and high serum urea nitrogen concentration, P less than 0.05) and prevalence of chronically diseased dogs was greater (P less than 0.005) in dog categories with higher U(P/C) values. More quantitative determinations of urinary protein loss as a screening test offer potential labor-saving and diagnostic advantages in the identification of urinary disease over more qualitative routine screening methods.