Why are you relying on meta-analysis of observational studies when we have meta-analysis of RCT's?