Abstract. Molecular biologic techniques are currently considered as new diagnostic and prognostic parameters with a sensitivity and specificity exceeding those of histologic and functional data currently used in clinical practice. The results in various clinical settings have been of limited value up to now. This study is an investigation of the use of tissue levels of RNA determined in routine clinical kidney biopsies as prognostic tools. The focus was on RNA encoding for molecules known to be involved in the pathogenesis of renal disorders. Fresh kidney biopsy tissue was obtained from 52 patients with various renal diseases. The GFR was followed for 12 mo. The extent of glomerulosclerosis and interstitial fibrosis in the biopsies was determined with quantitative digital image analysis. Glomerular and tubulointerstitial compartments from each biopsy specimen were separated, and mRNA levels of TGF-, collagen I, collagen IV, and fibronectin were quantitated by real-time PCR. Correlations, along with 95% confidence intervals (CI), between all variables tested at time biopsy were determined. To assess their prognostic value, these variables were correlated with the slope of GFR within several time intervals after biopsy. In addition, to evaluate the predictive value of the variables for outcome in individual patients, differences for each variable were tested between patients showing progressive decline in renal function (slope GFR Ͻ 0) and patients showing stable or improving renal function over time (slope GFR Ն 0). In chronic renal diseases, the extent of histologic damage correlated with the GFR at the time of biopsy (r ϭ Ϫ0.44; CI Ϫ0.68 to Ϫ0.11), but it did not correlate with the slope expressing a change in GFR after the biopsy. Tubulointerstitial TGF- mRNA levels correlated with the rate of change in GFR between time of biopsy and 1 mo later (r ϭ 0.41; CI, 0.07 to 0.67). The GFR at the time of biopsy correlated with the slope of change in GFR between time of biopsy and 12 mo later (r ϭ Ϫ0.50; CI, Ϫ0.73 to Ϫ0.18). In chronic renal diseases, glomerular fibronectin mRNA levels, in comparison with the GFR at time of biopsy, correlated relatively strongly with the slope of change in GFR between 3 and 12 mo (r ϭ 0.50; CI, 0.16 to 0.74). Patients with favorable renal outcome after 12 mo showed significantly higher TGF- mRNA levels and lower proteinuria levels at time of biopsy (P Ͻ 0.05) than patients with a progressive decline in renal function. This study shows that mRNA levels measured in kidney biopsies can function as prognostic tools in human renal diseases. In particular, relatively high levels of tubulointerstitial TGF- mRNA and glomerular fibronectin mRNA are associated with less deterioration in renal function.