Statins are widely prescribed for the prevention of vascular disease, based on a large body of evidence showing that they reduce mortality and the risk of vascular events (1). The idea that dyslipidaemia might promote progressive kidney disease was first advanced more than 100 years ago (2), and is supported by experimental and clinical evidence (3). It has been suggested that lipid-modifying treatment reduces pro- teinuria and prevents kidney function loss—especially statins that have been the most extensively studied class of agent for this indication. A recent systematic review of 27 randomized trials (with a total of 39 704 participants) suggested that statins reduce the rate of kidney function loss by 1.2 ml/min/year (95% CI 0.4, 2.0) or � 76% (4). A second systematic review of 15 randomized studies (1384 participants) found that statins reduce albumi- nuria by 47 (95% CI 26, 67) and 48% (95% CI 25, 71) in people with >300 mg/24 h and 30-300 mg/24 h of albumin excretion at baseline, respectively. However, statins did not significantly influence urinary albumin excretion when baseline levels were <30 mg/24 h (5). An article from the PREVEND group, in this issue of NDT, seeks to evaluate the effects of statins on the rate of change in estimated glomerular filtration rate (eGFR) and urinary albumin excretion. Atthobari et al. take the unconventional approach of presenting results from both an observational cohort study and a nested randomized controlled trial in the same paper. The randomized trial (PREVEND-IT) used a 2 � 2 factorial design to evaluate the effects of pravastatin 40 mg and/or fosinopril 20 mg daily compared with placebo in 864 normotensive people with microalbu- minuria (but not overt proteinuria), who were identi- fied by a screening program in the general population. The main results of this trial have been previously reported (7) and the current article focuses on the 392 subjects who did not receive fosinopril, among whom the rate of change in albumin excretion and eGFR were similar in pravastatin and placebo recipients. At first glance, the findings of PREVEND-IT appear to conflict with the meta-analyses described above. The discrepancy cannot be accounted for by dispara- ging the quality of the carefully conducted PREVEND- IT study, which was relatively large and had good internal validity, apart from a high frequency (25%) of loss to follow-up. Instead, differences in population characteristics may be responsible for the different findings compared with previous work. PREVEND-IT participants who received neither fosinopril nor statin had relatively preserved kidney function at baseline (mean eGFR 76 � 12 ml/min). These people were at low renal risk, since the average loss of eGFR was only 0.7 ml/min decline over 4 years, and few had clinically relevant declines in kidney function during the study period. This low event rate reduces statistical power to show an effect of statin on kidney function loss. No benefit of statins was noted in PREVEND-IT for urinary albumin excretion, which is interesting and worthy of publication. However, mean baseline albumin excretion was also very low in both