Monday, January 4, 2010

Alzheimer’s false scare

Icelandic company deCODEme, currently in financial straits, decided to offer clients of rival American company 23andme a nice Christmas present, allowing them to upload their 23andme raw genetic data to their website, at, in order to, a few hours later, and at no cost whatsoever, have them (re)analized through the deCODEme looking glass. What a fine idea, I thought – and took instantly advantage of their offer.

Two unpleasant things then happened. First, my mitochondrial haplogroup (maternal line) suddenly changed radically. According to 23andme, I’m H7, but at I was now K. Second, I was suddenly afforded the opportunity of knowing (but only apparently, as I have since learned) my lifelong-risk of getting Alzheimer’s Disease.

Concerning the first remarkable consequence of the deCODEme invite, almost immediately I received an email informing me that there had been a mistake in the interpretation of the mitochondrial data and apologizing for the inconvenience. I should wait a few days and everything would return to normal. By the way, I looked it up again this morning, and currently my mitochondrial haplogroup is back to H (simply H, which means it’s less informative than 23andme’s H7).

Concerning the second instance, I put off looking at my Alzheimer risk for days; I confess that I was afraid to see the results (you have to opt-in to access them). But this morning I learned that my inaction had probably been the best option – and for unexpected reasons: it turns out that, based on the 23andme results, deCODEme is unable to calculate that risk.

Only that didn’t keep some of those who had the audacity to opt-in to get really scary – and false – results! (see here). In short, the estimation of the risk is based on the detection of certain alterations in a gene called APOE, but 23andme doesn’t read the whole set needed to calculate it, which in this case makes the data useless – and sometimes exaggerated.

Currently, the opt-in option for Alzheimer’s has disappeared from my list of opt-in health conditions at But up till now, the company hasn’t offered any explanation for this, nor has it apologized for possibly scaring some people to death. I find this completely irresponsible.

Concerning myself, I learned something, though: I feel truly relieved not to have these results at hand for the time being, and I think that when the hour of truth comes, I may not have the courage to know it (the truth) with respect to certain particularly devastating diseases. I recently read a novel, Still Alice, by American neuroscientist-turned-writer Lisa Genova, about a case of early-onset Alzheimer’s. I manage to endure reality up to that point (I think the book is great), but if I get closer to it, I fear I might burn my wings...

Image: Drawing of neurons by Santiago Ramón y Cajal, 1899


  1. It is important to note that customers who buy a deCODEme complete scan receive an estimate of risk for Alzheimer's based on the SNPs rs429358 and rs7412 from the APOE locus. These are the most accurate markers to assess Alzheimer's risk due to APOE status. The chip used by 23andMe does not contain both of these SNPs. Consequently, the results of uploading 23andMe customers will be less comprehensive than those of deCODEme customers, as stated on the deCODEme upload page. As a result of this, risk for Alzheimer is not reported for this group.

  2. Alzheimer's Disease is a terrible disease that slowly takes our loved ones. It is imperative that we take a united stand against his terrible brain disease and come up with a cure.

    Alzheimer's Disease Support Canada