Eleven RCTs with a total of 56,308 patients were included.
In terms of study quality, in 7 studies the outcomes were assessed by blinded investigators; some other studies assessed outcomes by events committees.
ACEIs and ARBs significantly reduced the relative risk of developing AF compared with control by 28% (95% CI: 15, 40, p=0.0002). Statistically significant heterogeneity was found (p=0.0002).
AECIs and ARBs reduced AF by a similar amount: 28% (95% CI: 7, 44, p=0.01) for ACEIs and 29% (95% CI: 16, 40, p=0.0002) for ARBs. Statistically significant heterogeneity was found for both meta-analyses (both p<0.00001).
Treatment effects by medical condition.
Post MI (2 studies): the results were mixed. One study of patients with LV dysfunction found a significant reduction in AF with trandolapril for up to 4 years (RRR 0.52, 95% CI: 0.31, 0.87). The larger study, in patients with no heart failure, found no significant reduction in the risk of AF with 6 weeks of lisinopril compared with control (RRR 0.92, 95% CI: 0.83, 1.02).
Heart failure (4 studies): in patients with heart failure, ACEIs and ARBs significantly reduced the risk of AF by 44% (95% CI: 15, 63, p=0.007). All studies showed a significant reduction in AF but statistically significant heterogeneity was found (p=0.002).
In 3 studies the RRR appeared to increase with the severity of LV dysfunction: the RRR was 78% for 1 study of patients with severely impaired LV function (mean LVEF 26.7%), 23% for 1 study with mean LVEF 28%, and 18% for 1 study with mean LVEF 39%. The other study found a similar RRR in AF in patients with normal and impaired LV function.
Hypertension (3 studies): there was no significant reduction in AF with ACEIs and ARBs in patients with hypertension (RR 0.88, 95% CI: 0.66, 1.19, p=0.4). Statistically significant heterogeneity was found (p=0.001). Only the study of ARBs in patients with LV hypertrophy showed a significant reduction in the risk of AF.
Secondary prevention of AF after cardioversion (2 studies): most patients had hypertension and unimpaired LV function. There was a significantly reduction in AF with AECIs and ARBs in patients following cardioversion (RRR 48%, 95% CI: 21, 65). Neither study was placebo controlled.
Publication bias: the 5 smallest studies found the largest effect sizes but no negative small studies were found; this suggested the possibility of publication bias.