Background:​ This observational safety study used national registers to compare the real world cardiovascular and skeletal safety of zoledronic acid (ZA) against oral bisphosphonates (oBP) and untreated population controls.

Methods:​ Propensity score matched cohort study in Sweden and Denmark.

Results:​ Matched cohort 1 included 8739 ZA users and 25,577 oBP users while matched cohort 2 included 8731 ZA users and 25,924 untreated subjects.

In comparison to oBP users, heart failure risk was higher in ZA users, with an adjusted HR (adj) (95%CI) of 1.17 (1.04;​1.32) and a higher all-cause mortality (adj HR 1.24 (1.15;​ 1.34)), however, there was no increased risk of cardiovascular mortality. In the comparison to untreated subjects, ZA users showed a higher risk of atrial fibrillation, adj HR 1.18 (1.05;​1.32), arrhythmias adj HR 1.18 (1.06;​1.31), and heart failure, adj HR 1.38 (1.24;​1.54). Cardiovascular mortality was lower in ZA users (adj HR 0.87 (0.77;​ 0.98)) and risk of adverse skeletal outcomes was significantly higher, reflecting more severe osteoporosis in these patients. There was no association of cardiovascular risk with increasing exposure time. Sensitivity analyses produced similar findings with no substantial changes in event rates.

Conclusions:​ We noted an increased risk of heart failure, fractures and death among ZA users compared with oral BP. The risk of cardiovascular and skeletal outcomes was higher in ZA users than in matched population controls, but there was no increase in cardiovascular mortality in ZA users compared to oral BP or untreated controls. Despite propensity score matching, it is not possible to determine with certainty whether the increased risk of cardiovascular outcomes is consistent with a true drug effect or higher baseline risk in patients who begin ZA treatment.