I am working on a meta-analysis that examine the dose-response association of fasting plasma glucose (a continuous variable) on incident cardiovascular diseases. Although in most of the original studies HRs or RRs were reported, in some cases ORs were calculated that we should first convert ORs to RRs. Moreover, according to our aim, we want to pooled the adjusted effect sizes, not crude effect sizes. I know we have a formula for converting OR to RR in categorical variables that introduced by Zhang et al. However, for continuous variable I have not yet found any similar formula. Would you please introduce me any tips for my issue.