43Background: 44 With spot urine collected from a large control sample of preschool children (aged 3-7 years), 45 reference range of spot urine copper excretion indexes and their biological variation were defined.
47Methods:
48In order to investigate their test performance in screening of Wilson disease in this age group, 49 multiple spot urine samples from 6 WD patients diagnosed at presymptomatic stage were analysed. 50 Cut-off values for spot urine copper concentration, copper to creatinine ratio and copper to 51 osmolality ratio at 0.5 µmol/L, 0.1 µmol/mmol and 0.00085 µmol/mOsmol (32 µg/L, 56 µg/g 52 creatinine and 0.054 µg/mOsmol, respectively, in conventional units) have potential application in 53 differentiation of WD patients. 54 55 Results: 56 The data provides a new insight that the inter-individual variation of spot urine copper indexes 57 (CVg) were moderate with figures around 60% which was similar to other clinically useful urine 58 tests, such as urine albumin excretion ratio. Spot urine copper excretion strongly correlated with 59both urine creatinine and osmolality. And more than 95% of data points in health preschool children 60 fell within prediction regions by linear regression suggesting a good utility of normalisation by these 61 2 analytes. Receiver operator curve (ROC) showed that copper to osmolality ratio was the best 62 index with an area under curve (AUC) greater than 0.98. 63 64 109 important in the interpretation of all laboratory test results. Although large scale programs like 110 CALIPER have been set up to better define RI in the paediatric age group, they are confined to 111 analytes in serum or blood samples (12). It is reasonable as they are more commonly encountered in 112 routine laboratories.113 114 157 for controls. In addition, the prediction region covering the predicted distribution of 95% control 158 data points were shown (blue rectangles in scatter plots, e.g. figure 2). Difference in spot urine 159 indexes are compared between WD and control by non-parametric group-wise statistics (Wilcoxon 160 test). Statistical significance was defined by type I error of 0.05.161 162