Some days ago I wrote a letter to a researcher who is currently involved in the study of ME/CFS. I sent him some relatively rare genetic variants that I had found analyzing my own exome and the one of another patient (see this post). He was so kind to reply to my mail. He answered with a simple and – at the same time – very interesting note. If there was a genetic predisposition to ME/CFS – he observed – it would be common, very prevalent in the general population. Otherwise, we could not explain the epidemic episodes of the disease, like the one that happened in Lake Tahoe (Nevada), or in Lyndonville (New York), or in Bergen (Norway), and so forth. He left me with this problem “to think at night”, as he wrote.
Well, I did my homework. A genetic predisposition to ME/CFS has been suggested by a study on familial clustering of ME/CFS from a data bank of Utah health care system. They have found a significant increase in ME/CFS relative risk among first, second, and third degree relatives, compared with the general population (Albright F. et al. 2011). The problem is: acknowledged a genetic predisposition, how prevalent is it?
In 2004, a large outbreak of Giardia duodenalis struck the city of Bergen, in Norway. Of 1252 laboratory-confirmed cases, 347 reported chronic fatigue three years later. 53 of them were selected for a study and 41.5% of them were found to fit the criteria for ME/CFS (Mørch K et al. 2013). If we assume the same percentage for all the 347 patients who were symptomatic three years after the outbreak, we find that 144 of the original cases of laboratory-confirmed infection developed ME/CFS. This points to a prevalence of 11,5% in general population for the genetic predisposition to ME/CFS.
In order to confirm this result, I then considered a very well known study on the prevalence of ME/CFS among Australian patients who went to their doctor for infections due to Epstein-Barr virus, Coxiella burnetii, and Ross River virus. They found that after six months from the infection, 28/253 participants (11%) met the diagnostic criteria for ME/CFS (Hickie I. et al. 2006). So we find a slightly lower prevalence in this case, but we should consider that in the second study the diagnosis was made after six months from the infection, while in the first paper the clinical picture was evaluated after three years. The difference could be due to a certain degree of spontaneous recovery (4% of patients?) that has been reported (anectodal) during the first years from the beginning of the disease.
The conclusion of this very short and poor analysis is that if there was a genetic predisposition, it would be present in 11% of the general population. And yet, ME/CFS is much less prevalent. But if we consider the two studies mentioned, we could argue that we need a major infection (one that requires medical care and blood tests) in order to trigger this predisposition. So we would have a genetic predisposition highly prevalent (1 in 10 individuals!) but with low penetrance (only a small percentage of those who carry the genetic predisposition ends up developing the disease).
Now, if we assume that the genes involved in this predisposition are n and that these genes are transmitted independently one from another, then we have:
p_1 × p_2 × … × p_n = 0.11
where p_i is the prevalence of the variation on the i-th gene involved. This means that if we assume that the genetic predisposition is due to two or more genes, then each of these variants has a prevalence higher than 0.33.