Evaluation of semi-quantitative methods for protein and sugar estimation in urine

Abstract

To compare the accuracy of semi quantitative methods for estimation of protein and sugar in urine as shown by their agreement with the quantitative estimation. Hundred randomly collected samples of urine were analysed for levels of protein and sugar. Protein estimation was dine by dipstick and sulphosalicylic acid method (SSA) and sugar by dipstick and Benedict's semi-quantitative methods. Kappa analysis was done on Epi Info 6.03 software to assess the agreement of these semi quantitative methods with the quantitative estimation. Neither of the two tests for urine protein, dipstick or SSA, showed good agreement with the quantitative estimation (Kappa coefficient: 0.26 and 0.07 respectively). However, the dipstick was significantly better than SSA (p < 0.05). For urine sugar, both dipstick and Benedict's tests showed good agreement with the quantitative estimation (Kappa coefficient: 0.78 and 0.84 respectively). The difference between them was insignificant. Results demonstrate that for urine protein, dipstick or SSA show poor agreement with quantitative values. For urine sugar estimation, Benedict's semi-quantitative test shows good agreement with the quantitative values and is as good as the dipstick method.