Forty-nine trials (34,082 participants) were included in the review. Twenty-one trials were in patients with type 1 diabetes (5,413 were included in the review) and 29 in patients with type 2 diabetes (28,669 participants). Methods of randomisation were not reported for most of the trials.
Type 1 diabetes: In patients who were normoalbuminuric at baseline, the ratio of mean urinary excretion was not significantly different between treatment and control (0.94, 95% CI 0.79 to 1.12; Ι² not reported; seven studies; random-effects model). In patients who were microalbuminuric at baseline there was a significant effect in favour of treatment (0.33, 95% CI 0.23 to 0.46; 14 studies).
Type 2 diabetes: In patients who were normoalbuminuric at baseline the ratio of mean urinary excretion was significantly different between treatment and control in favour of treatment (0.79, 95% CI 0.68 to 0.93; seven studies). In patients who were microalbuminuric at baseline there was a significant effect in favour of treatment (0.73, 95% CI 0.62 to 0.85; 21 studies).
Results of subgroup analyses were presented in the paper and in an online appendix. For patients with type 1 diabetes, RAASI treatment of normoalbuminuric patients led to no significant difference in the number of patients who progressed to microalbuminuria from the comparator group. For patients with type 2 diabetes, RAASI treatment of normoalbuminuric patients led to significantly fewer patients who progressed to microalbuminuria (RR 0.84, 95% CI 0.79 to 0.89; Ι²=19%; eight studies).
In trials that included patients with type I and with type 2 diabetes, RAASI treatment of patients with microalbuminuria resulted in significantly fewer progressing to macroalbuminuria (RR 0.39, 95% CI 0.23 to 0.64; Ι²=0%; seven studies for type 1 and RR 0.52, 95% CI 0.43 to 0.63; Ι²=48%; eight studies for type 2) and in more patients regressing from microalbuminuria to normoalbuminuria (RR 5.81, 95% CI 2.05 to 16.43; Ι²=0%; four studies for type 1 and RR 1.20, 95% CI 1.12 to 1.29; Ι²=75%; eight studies for type 2).
There was no significant effect of treatment on mortality or glomerular filtration rate in any of the trials in patients with type 1 or type 2 diabetes.
Results of meta-regression were presented; no single factor could explain the heterogeneity between the trials.