I think I have something on my RG page dealing with metabolic syndrome. Before I specifically refer you for my data, I would like elaborate upon the fact that metabolic syndrome definitions have been evolving and varying over the years which can be easily depicted by seeing WHO, National Cholesterol Education Program(NCEP) and recent most the International Diabetic Federation (IDF) criteria for diagnosing metabolic syndrome.
Now coming to you question about fasting insulin levels in metabolic syndrome. Insulin levels have been seen not be helpful in isolation so various alternative strategies like direct measures including euglycemic clamp test have been devised but they are painfully difficult and not much in clinical practice. This led to appearance of various surrogate markers like HOMAIR, QUICKI and others to incorporate other factors like fasting glucose in helping detect resistance to insulin action in metabolic syndrome subjects.
I am posting some links from my articles on this subjects.
Article Insulin resistance in human subjects having impaired glucose...
I sincerely hope it should help. I also intend further on isulin in a while.
Further to your question on serum insulin variability, one has to understand that the variables affecting insulin causing variance to increase.
Firstly, beta cell insulin secretion is in response to insulin and early beta cell dysfunction can lead to higher surge in insulin which over time increases due to end organ insulin resistance.
Secondly, insulin is highly vulnerable to degradation, so quick storage and analysis is recommended.
Thirdly methodologies utilized to analyze insulin are immunoassays which range from plain manual ELISA to highly sophisticated platforms like chemiluminescence. Inter-platform variance is also recognized and the test is is in dire need of standardization.
Pre-analytical fasting, NAFLD, and inter-individual variability all contributes to increase CV for this parameter, and it probably this reason surrogate and direct measures of insulin incorporates stimulation techniques for use mathematical models like HOMAIR, HOMA%B and QUICKI which provides better information than insulin alone.