Obesity and Diabetes — 13

The graphs in figure 4.3 show the relationship of the polygenic risk score, by percentile, from zero to one hundred, for prevalence of type 2 diabetes in over three hundred thousand people in the UK Biobank (upper panel) and the odds ratio for developing type 2 diabetes among over two hundred thousand people in the Million Veterans Program (lower panel). Remember that type 2 diabetes is fundamentally different from its type 1 autoimmune diabetes counterpart, which is discussed in the chapter on autoimmune diseases.

The graphs in figure 4.3 are consistent and tell us that if one has a high score—90 or above—for the hundreds of genomic variants linked to type 2 diabetes, the risk of developing it is considerably increased. I’ve put both graphs in sequence to illustrate the difference between absolute risk (upper) and relative risk (lower). People with risk scores below 90 have only a small absolute risk, even though the odds ratio looks high. You can see how this absolute prevalence drops substantially below the 90 and 80 percent risk score (upper panel), whereas the big jump in relative risk is seen in the top 10 percent of polygenic risk score (lower panel).

Why is this important? For one, it helps explain why people who are not overweight or obese may still be at substantial increased risk for developing type 2 diabetes. For example, many thin people of Asian ancestry are prone due to their genetic makeup. About one of three people in the United States who are considered “pre-diabetic” and exhibit metabolic syndrome (vide infra), with abnormal fasting blood sugar (>110 mg/dl), are not obese. Whatever your body weight status, knowing your risk of diabetes can help promote its prevention, be it by changes in nutrition, exercise, and lifestyle, or with medications. Getting your polygenic risk score should not be expensive or difficult. The data can be derived from a gene chip (array) used by companies like 23andMe or AncestryDNA or low-pass genome sequencing. Either method detects the several hundred letter variants among the three billion letters of the human genome, and a formula is used to calculate the score, with weighting of specific variants by their importance, and critical attention to the person’s ancestry.