If you are only interested in comparing one analgesic against another, direct (i.e., head to head) comparisons provide the best information. You opt for indirect comparisons when there are no direct comparisons. Of course, you can combine the two, as in a network meta-analysis but I think most people would stick with direct comparisons. Any meta-analysis software can handle those.

I will recommend R software (R studio) if you want to compare multiple arms (compared to control/placebo). You can use bayesian methods or Frequentist method according to your data.

This article will definitely help you.

https://bookdown.org/MathiasHarrer/Doing_Meta_Analysis_in_R/netwma.html

In case you will compare only two drugs (head to head), you can use Review Manager 5.3 software.

Look at this paper

Article Incorporating multiple interventions in meta-analysis: An ev...