Twenty-seven RCTs (n=39,704) were included.
The studies were generally of a low quality. Four studies scored 4 or more points out of 5, while 16 studies scored 2 or less.
There was some asymmetry of the funnel plot suggesting the absence of a few small negative trials. The weighted regression test showed no evidence of publication bias (p=0.24).
Effect of statins on GFR.
Statins were associated with a statistically significant reduction in the rate of change in eGFR compared with control (WMD 1.22 mL/minute per year, 95% CI: 0.44, 2.00). There was substantial heterogeneity (I-squared 96%).
Statins were associated with a statistically significant reduction in the rate of loss of renal function in patients with CVD (eGFR 0.93 mL/minute per year, 95% CI: 0.10, 1.76); the substantial heterogeneity persisted (I-squared 99%; based on 6 cohorts, n=38,311). There were no significant differences between statins and control in change in eGFR for patients with diabetes (based on 5 cohorts, n=122), GN (based on 7 studies, n=222), HTN (4 cohorts, n=212), or all these subgroups combined. The results were similar when only studies lasting 1 year or more were analysed.
The mean age, mean total serum cholesterol, mean change in baseline serum cholesterol in the statin treatment group, baseline eGFR and population (CVD disease versus non-CVD) were not significantly associated with changes in eGFR. Adjusting for atorvastatin (compared with other statins) reduced the level of heterogeneity (from I-squared 96% to 46%).
The effect of statins remained statistically significant with a smaller treatment effect when only higher quality studies were analysed.
Effect of statins on proteinuria and albuminuria.
There were no significant differences between statins and control in the rate of change in proteinuria (reduction in WMD 0.37 g/24 hours, 95% CI: -0.75, 0.02) or albuminuria (reduction in WMD -0.02 g/24 hours, 95% CI: -0.06, 0.02); substantial heterogeneity was found (I-squared 83% and 92%, respectively).