Stimulated by Lv et al 2023.[1]
MDD – major depressive disorder
key to acronyms
TCM – traditional Chinese medicine
GWAS – genome wide association studies
SNP – single nucleotide polymorphism (change of a single letter in the genetic code)
PD – Parkinson’s disease
CHD – coronary heart disease
This paper came up on my searches last month, but it does not concern either myofascial pain or acupuncture. Increasingly, papers come up because acupuncture is mentioned in the affiliations of some Chinese authors from TCM hospitals and universities.
MDD and migraine are both conditions that as acupuncture practitioners we might legitimately try to help; however, this research falls into the category of genomic epidemiology rather than therapeutics and I chose it because the term Mendelian randomisation was entirely new to me, despite having hit the epidemiological literature 20 years ago.[2]
Now whilst Gregor Mendel (and hence Mendelian) rings some bells when it comes to pea shoots and inheritance of recessive and dominant traits, and randomisation is a very familiar term, the combination of the two to form Mendelian randomisation took me by surprise.
It is a fairly simple idea based on the random assortment of genes delivered from parental gametes to their offspring. So, if you can find the genetic basis for a certain risk factor for some disease using the ever-increasing GWAS data,[3] you can go on to compare groups based on whether or not they have that genetic profile (usually a set of SNPs) without the risk of selection bias, since the distribution of genes is random in the population. You also avoid the problem of reverse causality – mistaking a factor associated with a disease as the cause of the disease when it is actually the result of the disease instead, and therefore not worth trying to alter because it is downstream in the disease process. Genes are not altered by disease processes in the host, so there is no risk of reverse causality when using randomisation based on the genome.
In this case, the risk factor is MDD or migraine, and the outcome is migraine or MDD. So, GWAS data provides both the genetic variants associated with the conditions, but also the essentially random samples of the population with respect to the outcome condition.
There are clearly important assumptions here. Mendelian randomisation assumes that the genetic variant (the group of SNPs chosen by the team) is associated with the risk factor, in this case the first condition (MDD or migraine). It also assumes that the same genetic variant is not associated with any other conditions that might influence the outcome disease (confounders). Lastly, it also assumes that the genetic variant identified only influences the outcome disease via the risk factor of interest (in this case the first condition, ie MDD or migraine).[4]
So, when looking for a causal relationship between MDD and migraine, you take GWAS data concerning MDD, including both cases and controls, and interrogate it to find the prevalence of the migraine phenotypes. And, when looking for a causal relationship between migraine and MDD, you take GWAS data concerning migraine and interrogate it to find the prevalence of MDD.
This may sound straightforward, but we are not dealing with black and white associations, and we are dealing with massive data sets. Each SNP or group of SNPs predicts the risk factor disease to a variable extent, and then we have to counter pleiotropy and linkage disequilibrium… what are you on about now Mike!
Pleiotropy refers to the genetic variant influencing the outcome (migraine or MDD here) independent of the risk factor of interest (MDD or migraine respectively).[4]
Linkage disequilibrium refers to the non-random association of alleles at different loci on the genome.[5] Yes, the very thing that is assumed not to occur for Mendelian randomisation to be valid in the first place.
The process of meiosis (cell division to form gametes) is not truly random in terms of genes because the process does not randomly divide each gene, so loci of the genome that are close to each other in the parent DNA are more likely to stay together in the gamete than loci that are distant from each other.
This effect can be reduced by excluding SNPs that do not have a big enough effect on the risk factor and may influence the outcome. I think I will leave it there… it starts to get really complex when you try to address pleiotropy.
The bottom line is that from this analysis MDD is a risk factor for migraine, but migraine is not a risk factor for MDD.
We have already seen that the risk of stroke, PD, and CHD in patients with MDD is reduced in association with the patients receiving acupuncture in Taiwan.
References
1 Lv X, Xu B, Tang X, et al. The relationship between major depression and migraine: A bidirectional two-sample Mendelian randomization study. Front Neurol 2023;14:1143060. doi:10.3389/fneur.2023.1143060
2 Smith GD, Ebrahim S. “Mendelian randomization”: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:1–22. doi:10.1093/ije/dyg070
3 Uffelmann E, Huang QQ, Munung NS, et al. Genome-wide association studies. Nat Rev Methods Primer 2021;1:59. doi:10.1038/s43586-021-00056-9
4 Emdin CA, Khera AV, Kathiresan S. Mendelian Randomization. JAMA 2017;318:1925–6. doi:10.1001/jama.2017.17219
5 Slatkin M. Linkage disequilibrium–understanding the evolutionary past and mapping the medical future. Nat Rev Genet 2008;9:477–85. doi:10.1038/nrg2361
The BMAS Blog 
You must be logged in to post a comment.