Many studies have investigated the association between stroke and hip fracture risk, but the precise association was still unclear due to insufficient statistical power in single studies with relatively small sample size. Thus, we firstly conducted a meta-analysis of all published studies to precisely estimate the relationship of stroke with hip fracture risk. The strength for this relationship was weighed by pooled relative risks (RRs) with 95 % confidence intervals (95 % CIs) after adjustment for confounding variables. Stratified analyses by study design and ethnicity and sensitivity analysis were also performed. Two investigators independently performed a comprehensive literature search in databases of PubMed, Embase, and Wanfang for eligible articles. A Bayesian meta-analysis was also performed to get a more precise assessment of the relationship. Eleven relevant studies from 10 publications were finally included into our meta-analysis according to the inclusion criteria. Overall, stroke significantly and independently increased the risk of hip fracture (RR = 2.06, 95 % CI 1.68-2.52, P < 0.001). Bayesian meta-analysis showed that stroke was also associated with an over two-fold increased risk of hip fracture (RR = 2.11, 95 % CI 1.62-2.75). In stratified analysis, stroke could increase the risk of hip fracture in Caucasians (RR = 2.36, 95 % CI 1.83-3.05, P < 0.001). These data support the notion that stroke is an independent risk factor for hip fracture, and patients with stroke have a two-fold increased risk of hip fracture than those without stroke.
Keywords: Cardiovascular disease; Hip fracture; Stroke.