Four RCTs (26,342 participants) and three observational studies (130,710 participants) were included in the review. The quality of RCT and observational studies were both deemed good. Trials were double-blinded. Trial duration ranged from 22 to 50 months.
The RCTs showed a statistically significantly greater risk of atrial fibrillation with bisphosphonates compared with placebo (RR 1.525, 95% CI 1.166 to 1.997, I2 = 53%; four RCTs).
Two out of three observational studies showed a statistically significant increase in the risk of atrial fibrillation with bisphosphonate therapy. One study showed no difference. There was some evidence that alendronate had a greater risk of atrial fibrillation compared with etidronate (one study).