You need to standardize these different forms into a single form to run your meta-analysis. For instance, you can convert continuous data in model 2 into classes like model 1. As for model 3, they probably had a single group i.e mean BMI was

Dear Mohamed Abdel-Maboud ,

Thank you very much for your answer. As I am trying to do a linear dose-response meta-analysis, I think converting number (2) to number (1) seems inappropriate; rather, I'm trying to convert number (1) to number (2).

The main question is:

Can I interpret number (3) the same way with the interpretation of number (2)?

I believe that the example number (2) fits the linear dose-response meta-analysis well. However, I am still clueless whether example number (3) can be interpreted same to example number (2). If they can, example number (3) can also fit the linear dose-response meta-analysis. Otherwise, I will have to exclude them from the final model.

Thank you very much for the help.

Best,

If you can categorize or assign their reported data into classes like model 1 i.e =30 kg/m2 then you can include them. Otherwise, you may exclude them from the analysis and highlight this in the discussion.