Evidence for a Causal Association Between Human [CMV] Infection and Chronic Back Pain: A One‐Sample Mendelian Randomization Study, 2025, Naeini et al

[forestglip]

Okay yes, two-stage least squares (2SLS) regression does work by estimating the causal effect as a ratio between the effect of the allele on the exposure and the effect on the outcome.

So in effect, you're starting out by looking all the alleles which are associated with CMV from the GWAS, and then regressing out the degree to which those same alleles are predictive of back pain on their own. What you're left with is the degree to which the allele is predictive of back pain only when there was CMV infection.

It looks like linkage disequilibrium and horizontal pleiotrophy are the two biggest possible confounders here--they tested for LD, but didn't seem to do anything additional for the latter. I'm not enough of an expert to determine if there was justification to skip that in this case.

Either way, it seems like examples of more robust analyses typically include the MR-Egger intercept test, as @forestglip brings up.

Here's some additional sources I looked at to understand the 2SLS regression method better:
Mendelian Randomization as an Approach to Assess Causality Using Observational Data
Basic Concepts of a Mendelian Randomization Approach

Thanks for the info. I spent some time trying to figure out what exactly two stage least squares regression means and found I was out of my depth with all the explanations.

 

[jnmaciuch]

Thanks for the info. I spent some time trying to figure out what exactly two stage least squares regression means and found I was out of my depth with all the explanations.

I sympathize, I think I barely understand it myself and that's only because I spent a lot of time learning the terminology of regression analyses. The math has been around long enough that I'm pretty confident it's correct (I don't have the skill to check myself, but it's been subjected to a lot of scrutiny). So the issue would just be knowing how confident you can be about the assumptions of MR in general