Fourteen RCTs were included (1,093 participants; 1,179 fractures). Nine studies reported concealment of allocation using sealed envelopes. Full blinding was not possible as calcium phosphate cement is radiodense. There were imbalances between groups at baseline in four studies. Three studies did not report a intention-to-treat analysis. On the Detsky scale assessing quality of reporting, scores ranged from 50 per cent to 90 per cent.
Three (n=455 patients) of the eight studies reporting a pain outcome were pooled. The prevalence of pain was lower at the fracture site in the intervention group (calcium bone cement) compared to control (RR 0.57, 95% CI: 0.33, 0.99, p=0.04). In the subgroup analysis there was no statistically significant difference in pain when the intervention was compared to autogenous bone graft (one RCT, n=22), but there was a significant pain benefit when the intervention was compared to a no-substitute comparator (two RCTs, n=432).
Appropriate data were not available for pooling of the five studies reporting functional outcome. The findings from these five studies were mixed.
There was no significant difference between intervention and control prevalence of infection (seven RCTs, n=718), RR 0.74 (95% CI: 0.19, 2.87), in fracture healing (four RCTs, n=216) and in prevalence of loss of fracture reduction (five RCTs, n=599), RR 0.54, 95% CI: 0.26, 1.13). There was significant heterogeneity. When the intervention was compared to autogenous bone graft alone in a subgroup analysis there was a statistically significant benefit in favour of the intervention (three RCTs, n=166) for loss of fracture reduction, but not when compared to no substitute. There was a benefit with the intervention for tibia fractures (two RCTs, n=146), but not with other fractures. The authors stated that funnel plots did not suggest publication bias.