Analysis of a urinary biomarker panel for incident kidney disease and clinical outcomes

Abstract

Whether novel biomarkers improve the assessment of incident kidney disease and related adverse outcomes remains to be tested in longitudinal observational studies. We tested 14 urinary biomarkers for association with incident kidney, cardiovascular, and mortality outcomes in 2948 Framingham Heart Study participants. Baseline examinations were performed between 1995 and 1998; mean follow-up was 10.1 years for renal outcomes and 11.2 years for survival analyses. Primary outcomes were incident CKD, incident albuminuria, incident cardiovascular disease, and all-cause mortality. Secondary analyses assessed incident congestive heart failure (CHF) and mortality with coexistent kidney disease. Biomarkers were tested for association with renal end points using logistic regression and incident cardiovascular and mortality outcomes in proportional hazards models; α1-microglobulin, Kim-1, and TFF-3 predicted all-cause mortality (hazard ratio per SD increase in log-transformed biomarker [HR] range, 1.15 to 1.21; 95% confidence interval [CI] range, 1.04 to 1.34; P values=0.007 to <0.001), whereas α1-microglobulin, β2-microglobulin, KIM-1, and TFF-3 associated with death with coexistent kidney disease (HR range, 1.72-2.25; 95% CI, 1.17 to 3.24; P values<0.01). KIM-1 also associated with the risk of incident CHF (HR, 1.32; 95% CI, 1.07 to 1.63; P=0.008). CTGF associated nominally with CKD (HR, 0.83; 95% CI, 0.71 to 0.98; P=0.03), but no other biomarkers associated with incident CKD or albuminuria. Addition of α1-microglobulin and TFF-3 resulted in a nonsignificant net reclassification index (NRI) of 3% for all-cause mortality beyond clinical risk factors. In conclusion, components of a panel of 14 subclinical biomarkers of kidney injury were associated with important clinical outcomes and merit additional investigation.

Figure 1.

Mortality rates increase by quartile of α1-microglobulin, KIM-1, and TFF-3 in participants with and without CKD. Bars represent age- and sex-adjusted mortality rates, and error bars represent standard errors quartiles of α1-microglobulin, KIM-1, and TFF-3.