Figure 1.
Creatinine, Osmolarity and Protein distributions.
Triplicate levels of the standards were measured in 120 hematuric patients and then averaged. (A) Osmolarity was normally distributed; (B) creatinine and (C) protein had skewed distributions.
Figure 2.
Relationship between osmolarity and creatinine.
Triplicate levels of osmolarity and creatinine were measured in urine from 119 hematuric patients. There was a modest relationship between osmolarity and creatinine (R Square = 0.519).
Figure 3.
Comparison between measured protein levels and protein dipstick analyses.
Total protein levels (mg/ml) in urine were determined by Bradford assay A595 nm (Hitachi U2800 spectrophotometer) using Bovine Serum Albumin as standard. Dipstick analyses were undertaken using Aution Sticks 10EA. Analyses were interpreted using PocketChem (Arkray factory, Inc. Japan). Protein levels were plotted against dipstick results with the Y –axis reference line indicating the usual lower limit of sensitivity for urine dipstick testing (0.25 mg/ml).
Table 1.
Comparison of protein levels across final diagnostic categories.
Figure 4.
AUROC for IL-6, IL-8 and VEGF.
The lowest area under receiver operating characteristic (AUROC) were determined after protein normalization as represented by the solid black curve which was always closest to the diagonal reference line i.e., IL-6 = 0.634 (0.523 to 0.745); IL-8 = 0.677 (0.570 to 0.784); and VEGF = 0.609 (0.501 to 0.716). The AUROCs for uncorrected biomarker levels (thick grey curve), and those standardized using osmolarity (dashed black curve) or creatinine (dashed grey curve) were very similar for individual biomarkers : (A) IL-6 = 0.693 (0.592 to 0.794), 0.683 (0.582 to 0.784) and 0.678 (0.578 to 0.779), respectively; (B) IL-8 = 0.706 (0.608 to 0.804), 0.701 (0.603 to 0.799) and 0.694 (0.592 to 0.795), respectively; and (C) VEGF = 0.705 (0.610 to 0.799), 0.687 (0.591 to 0.783) and 0.680 (0.583 to 0.777), respectively. Figures in brackets are 95% Confidence Intervals.
Figure 5.
Bland Altman plots for osmolarity, creatinine and protein.
Bland Altman plots for (A) osmolarity; (B) creatinine; and (C) protein (loge) were plotted to determine the agreement between the levels of each standard measured on recruitment and those measured at follow-up. The hashed line (mean of the mean differences) demonstrates that protein levels decreased by approximately 24% at follow-up. Osmolarity and creatinine levels did not significantly change. Solid lines, 95% CI limits. Open triangles (bladder cancer); closed black circles (controls).
Table 2.
Paired t-test comparing standard levels measured on recruitment and at follow-up.
Figure 6.
Regression analyses to determine the relationship between differences in standards and biomarkers over time.
Scatter plots, based on data from 72 hematuric patients, plotting the differences between biomarker levels on recruitment and follow-up against the differences between protein levels on recruitment and follow-up for (A) IL-6, (B) IL-8 and (C) VEGF. The regression line and 95% confidence interval show significant associations (p<0.0001 for all biomarkers). Differences in biomarker levels across time were associated with differences in protein levels.
Figure 7.
SDS PAGE was carried out on urine from each patient. A dense band was frequently observed at approximately 64–66 kDa. This band represents albumin. Eight representative samples demonstrate the diverse relationship between this albumin band on the SDS PAGE and corresponding IL-8 levels measured in urine from the same patient sample. Corresponding IL-8 levels are illustrated in the 95% confidence limit error bar chart directly below each lane. The density of the albumin band was not always indicative of the IL-8 levels. Four patients had non-muscle invasive bladder cancer (NMI), one patient had muscle invasive bladder cancer (MI), two patients had no diagnosis (ND), and one patient had benign prostate enlargement.