UNLABELLED

Benzodiazepines (BZDs) are some of the most commonly prescribed drugs in the world. It has been shown that BZD use could be associated with increased fracture risk. However, studies on the use of BZDs and fracture risk have yielded inconsistent results. Results from the present meta-analysis show that BZD use is associated with a moderate and clinically significant increase in the risk of fractures.

INTRODUCTION

The relationship between the use of BZDs and fracture risk has been neither well identified nor summarized. This meta-analysis reports on the use of BZDs, especially short-acting BZDs, and their correlation with a moderate and clinically significant increase in fracture risk. This analysis will provide evidence for clinicians to consider fracture risk when prescribing BZDs among the elderly population. This study was conducted to determine whether people who take BZDs are at an increased fracture risk.

METHODS

A systematic search of studies published through January 2013 was conducted using MEDLINE, EMBASE, OVID, and ScienceDirect. Case-control and cohort studies that assessed the relationship between BZD use and the risk of fractures were identified. Literature searches, study selections, methodological assessments, and data mining were independently conducted by two reviewers. Disagreements were resolved by consensus. STATA 12.0 software was used for the meta-analysis. Random effects models were used for pooled analysis due to heterogeneity among the studies.

RESULTS

There were 25 studies, including 19 case-control studies and 6 cohort studies, that met the inclusion criteria. Overall, the results of the meta-analysis indicated that BZD use was associated with a significantly increased fracture risk (relative risk (RR) = 1.25; 95% confidence intervals (CI), 1.17-1.34; p < 0.001). Increased fracture risk associated with BZD use was observed in participants aged ≥65 years old (RR = 1.26; 95% CI, 1.15-1.38; p < 0.001). When only hip fractures were included as the outcome measure, the RR increased to 1.35. However, subgroup meta-analyses showed that there was no significant association between BZD use and fracture risk in Eastern countries (RR = 1.27; 95% CI, 0.76-2.14; p = 0.362) as well as between long-acting BZD use and risk of fractures (RR = 1.21; 95% CI, 0.95-1.54; p = 0.12). After accounting for publication bias, we observed that the overall association between BZD use and fracture risk to be slightly weaker (RR = 1.21; 95% CI, 1.13-1.30) but still significant.

CONCLUSION

The results of this meta-analysis demonstrate that the use of BZD, especially short-acting BZD, is associated with a moderate and clinically significant increase in fracture risk. However, large prospective studies that minimize selection bias are necessary to determine a more accurate fracture risk associated with BZD use.