2016-03-13_10:27 by Michael Koren
Diabetes certainly seems to be a much more heterogeneous disease than the current classification would suggest. Stan's classification is practical. Joel Dudley used an interesting big data mining approach in identifying subgroups within type 2 diabetes patient population (http://stm.sciencemag.org/content/7/311/311ra174). MODY, especially de novo mutations, probably significantly underrecognized. To answer your questions:
1. About 5% sounds right
2. When in doubt about whether the patient needs to stay on or to switch to insulin I am measuring C-peptide and GAD 65 first. If GAD 65 is negative, but suspicion for an autoimmune diabetes is strong, I would also check IA2 and insulin Ab. We use Mayo Lab as our reference lab and they do not have ZnT8 in their catalog.
3. I, as many of you I am sure, see a fair number of obese adults diagnosed with type 1 as children. Considering large prevalence of obesity in the population I usually do not check their antibody status unless they admit/report their non-adherence with insulin without the ensuing DKA.
4. About 5-10%
