Thirteen studies (319,148 patients) were included in the review, comprising seven cohort studies (130,637 patients) and six case control studies (188,511 patients). The results of the quality assessment were not presented, although the results of the subgroup analyses showed that 11 studies attained quality scores of 7 points or higher. Duration of follow-up ranged from four years to 8.4 years.
There were significantly higher risks of fractures observed for patients who received selective serotonin re-uptake inhibitors compared with patients with no exposure (RR 1.72, 95% CI 1.51 to 1.95; 12 studies). The fracture risk remained statistically significant for studies that controlled for important risk factors (RR 1.70, 95% CI 1.48 to 1.94; 10 studies), studies that included patients aged 50 years or older (RR 1.68, 95% CI 1.41 to 2.00; nine studies), studies that used non-vertebral fractures as an outcome (RR 1.78, 95% CI 1.43 to 2.21; eight studies), studies that used hip fracture as an outcome (RR 1.70, 95% CI 1.48 to 1.95; seven studies), and studies that scored 7 or higher in the quality assessment (RR 1.63, 95% CI 1.46 to 1.81; 11 studies).
There were no significant differences between patients who received selective serotonin re-uptake inhibitors and those who did not in bone mineral density as measured by bone loss at the hip (0.19%, 95% CI -0.15 to 0.53; two studies).
The Cochran Q statistic indicated significant heterogeneity (p<0.001).
There was some potential for publication bias as observed by the funnel plot and the Begg test, although after the trim-and-fill analysis correcting for missing data, the risk of fracture was still statistically significant (RR 1.40, 95% CI 1.22 to 1.61)