All recent studies of the outcome of different forms of progressive glomerulonephritis concur that a major factor, apparently determining outcome, is the presence and severity of tubulointerstitial changes, and not the degree of glomerular alteration. Moreover, at the time of biopsy, tubulointerstitial changes correlate much better with the glomerular filtration rate. These at first surprising findings are not only useful clinically, but should make us think about our models of how progression takes place in so-called glomerular nephritides. In fact, a major tubulointerstitial infiltrate of immune-competent cells is present in all forms of progressive glomerulonephritis, and again correlates with outcome. In addition, it is now clear the tubular epithelium is capable of synthesising and secreting a number of factors important in fibrogenesis, and of displaying major histocompatibility complex class II antigens and leucocyte-adhesion molecules. Tubular cells could thus present peptides to T helper cells and amplify, or maybe even initiate, immune reactions. Finally, fibrogenesis within the kidney is at last being studied, long after studies have been performed on liver and lung. In the past, too much attention has been paid to reversible inflammation and not enough to irreversible cirrhosis of the kidney.