Performance indicators and validity of different analytical methods for measuring urine protein and microalbumin in patients with diabetes mellitus

Abstract

Objectives: 1. To derive a reference range for random urine protein:creatinine index (PCI) and albumin: creatinine index (ACI) in apparently healthy subjects, using different analytical methods. 2. To compare and assess the validity of sulphosalicylic acid (SSA) and pyrogallol red (PGR) methods for measuring urine protein as alternatives to immunoturbidimetric (IT) method for measuring urine microalbumin, in Iraqi diabetics with or without microalbuminuria (MA) or/and proteinuria.
Subjects and Methods: Random urine and fasting blood specimens were collected from 400 diabetics (256 females, 144 males) aged 8-87 years, including 48 type 1 and 352 type 2 diabetics. They were attending Al-Waffa Diabetic Clinic in Mosul during 6 months from 1st August 2002 to 31st January 2003. A control group of 145 apparently healthy volunteers (108 females, 37 males) aged 15-72 years were used for comparison. Urine protein was measured using SSA and PGR (for all diabetics and controls) and urine albumin using IT (for 112 diabetics and 75 controls). The statistical methods used included unpaired student Z-test and linear regression analysis. The validity indicators: sensitivity, specificity, negative and positive predictive values and accuracy rate were calculated.
Results: The frequency distribution of PCI and ACI showed log-normal distribution and following log transformation, the reference range for PCI was 20-235 mg/g using SSA and 18-205 mg/g using PGR, and for ACI was 4-55 mg/g using IT. The overall prevalence of proteinuria in the diabetics was 30% using SSA method and 35% using PGR method and MA was 27%. The SSA and PGR methods for measuring proteinuria were compared with IT method for measuring MA. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy rates were 93%, 96%, 90%, 98% and 95.5% respectively for PGR; and 80%, 95.1%, 86%, 93% and 91% respectively for SSA. A highly significant correlation (P < 0.001) was observed between ACI and PCI values (r = 0.85 using PGR, r = 0.88 using SSA) in all control and diabetic subjects. Prediction of ACI from PCI value can be made by multiplying the PCI value by 0.375 using PGR and 0.39 using SSA.
Conclusion: Proteinuria and MA are common among Iraqi diabetics. Simple and cheap methods, particularly the PGR method, have acceptable performance to be routinely implemented in diabetic care. It is recommended to measure random urine PCI in all diabetics during their regular visits to the diabetic clinic.
Keywords: Microalbuminuria, sulphosalicylic acid, pyrogallol red.