Effects of β-carotene intake on the risk of fracture: a Bayesian meta-analysis

Abstract

Background: Epidemiological studies examining the association between β-carotene intake and risk of fracture have reported inconsistent findings. We conducted a meta-analysis to investigate the association between β-carotene intake and risk of fracture. Methods: We systematically searched PubMed, EMBASE and Cochrane library databases for relevant articles that were published until December 2019. We also identified studies from reference lists of articles identified from the clinical databases. The frequentist and Bayesian random-effects model was used to synthesize data. Results: Nine studies with a total of 190,545 men and women, with an average age of 59.8 years, were included in this meta-analysis. For β-carotene intake (1.76–14.30 mg/day), the pooled risk ratio (RR) of any fracture was 0.67 (95% Credible Interval (CrI): 0.51–0.82; heterogeneity: P = 0.66, I2 = 0.00%) and 0.63 (95%CrI: 0.44–0. 82) for hip fracture. By study design, the pooled RRs were 0.55 (95% CrI: 0.14–0.96) for case-control studies and 0.82 (95% CrI: 0.58–0.99) for cohort studies. By geographic region, the pooled RRs were 0.58 (95% CrI: 0.28–0.89), 0.86 (95% CrI: 0.35–0.1.37), and 0.91(95% CrI: 0.75–1.00) for studies conducted in China, the United States, and Europe, respectively. By sex, the pooled RRs were 0.88 (95% CrI: 0.73–0.99) for males and 0.76 (95% CrI, 0.44–1.07) for females. There was a 95% probability that β-carotene intake reduces risk of hip fracture and any type of fracture by more than 20%. Conclusions: The present meta-analysis suggests that β-carotene intake was inversely associated with fracture risk, which was consistently observed for case-control and cohort studies. Randomized controlled trials are warranted to confirm this relationship.