The natural history of membranous nephropathy (MGN) was first described in the late 1960s by Pollock et al. (1). This paper clearly identified that only 40%-50% of patients will progress over time and up to 30% will have a spontaneous remission. This natural history has both vexed and fascinated clinical nephrologists since that time.All nephrologists can appreciate the value of early separation of progressors from nonprogressors given our current potent immunosuppressive drugs with their significant adverse effects. Conservative therapy alone has its own downsides, including an increased thromboembolic risk, premature atherosclerosis, and the potential for irreversible kidney injury (2,3). Recent data confirmed the significant spontaneous remission rate, regardless of initial proteinuria and, in some cases, regardless of the initial renal function (4,5).Almost two decades ago, the Toronto group looked at whether the persistence of proteinuria above certain cutoff values would improve the overall accuracy of predicting chronic renal insufficiency. When we added initial renal function and rate of renal function decline over a time frame of observation to these proteinuria cutoffs, we improved baseline prediction from 25% to an accuracy of 80% (6). However, there are challenges with this prediction tool, including an observation period of up to 18 months and a somewhat complex calculation to determine the risk score.Many advances in nephrology practice have been made since that time. Significant improvements in conservative management of the proteinuric patient include widespread use of renin-angiotensin systemblocking agents and lower target BP. In addition, there are now new options in immunosuppressive drugs. This is on a background of changing demographics with the average presenting MGN patient now older by two decades and many with additional comorbidities such as hypertension and obesity compared with the original cohort (7). All these changes bring into sharp focus the question of whether the old methods of prediction still apply. This is a question addressed in an article by van den Brand et al. (8) on other urinary markers in this issue of CJASN. This different approach to prediction has been championed for over a decade by this group examining MGN outcome relative to the initial levels of low and high molecular urine protein biomarkers (9-12). The current article is the first to compare predictive accuracy of these two approaches in MGN.The authors conclude that the area under the receiver operating characteristic curves (ROC-AUC) between the Toronto risk score and either of their urinary protein biomarkers, b2 microglobulin or urinary a1 microglobulin are very similar despite exposure to the newer conservative treatment agents including angiotensin-converting enzyme and/or angiotensin II receptor blocker therapies (ROC-AUC 0.78 versus 0.795; 95% CI, 0.69-0.88). This conclusion must be tempered in light of the limited details about renin angiotensin system blockade exposure because neither dose nor dur...