Is Type I diabetes an evolutionary adoptation?

At least one evolutionary biologist thinks so. According to this NYTimes report, Dr. Sharon Moalem proposes that Type I diabetes “evolved” in Northern Europe during a period of sudden increase in cold weather 14,000 years ago. The paper is here (full text only with subscription, as usual. Note the ridiculous $30 to read the article).

The basic idea is that increases in glucose and various alcohols associated with type I diabetes in the blood may be protective against freezing of tissue during extreme cold. Type I diabetes is most common in northern European populations, Therefore, type I diabetes was selected for by cold exposure. They throw in a bunch of mumbo jumbo about ice ages and adaptations of other animals, but this doesn't advance their argument much

Let me first say I applaud the authors for trying to find an evolutionary explanation for this disease. This approach may be misguided, but I like this kind of thinking.

That said, the theory is wildly speculative.

First, it is not clear to me that type I diabetes evolved. More likely, it is an unfortunate side effect of an effective immune system. Evolutionarily, there is a trade off between an immune system that can effectively fight off infections and autoimmunity (the “horror autotoxicus” as Paul Ehrlich termed it). Type I diabetes indubitably results from immune system mediated destruction of insulin producing beta cells in the pancreas. Thus, type I DM, and other autoimmune diseases, may simply be a result of a vigorous immune system. Given how rare type 1 DM is, it doesn’t seem an unreasonable price for the ability to fight off infections, particularly given how common those were as a cause of death until 60 years or so ago.

Second, Type 1 DM is a rare disease today, and it was probably much rarer 10,000 years ago, as incidence has dramatically increased in the last few decades In a figure in the article, the highest prevalence is <40/100,000/year. Assuming about people are at risk for ~10 years you get a lifetime incidence of about 1/2000. I’m not an evolutionary theorist, but that seems a low frequency for something that provides a significant evolutionary advantage.

Third, having actually taken care of kids with type I diabetes, I find it hard to believe they could ever have a selective advantage. They are very sick and die if they don’t get insulin, which obviously wasn’t available 10,000+ years ago. In the article they hand wave about limited life spans at that time, but the life span of an untreated diabetic is generally very short. In addition, poorly controlled blood sugars during pregnancy are linked to high rates of fetal defects and trouble with blood sugars at birth (they have very high insulin levels in reaction to all the glucose they get from mom in utero), so I doubt the offspring of untreated female diabetics would have very good survival.

Finally, if anything I’d think type 2 diabetes would be selected for. After all, type 2 diabetics have high blood sugars and less severe forms of many of the metabolic derangements seen in type 1. But the disease progresses much less slowly and they are much less sick.

To summarize while I approve of generating interesting hypothesis, I find this one implausible.