There is an excellent review of the diagnosis of diabetes in the current issue of the New England Journal of Medicine (Vol 367, pp 542-550, Aug. 9, 2012) so I thought I would summarize it here and add a few relevant. clinical facts as well.
First, except for pregnant women, there is no good clinical evidence that screening and early treatment of "pre-diabetes" improves any clinical outcome. Two different tests are used to screen for diabetes and prediabetes: the fasting (8 hours) glucose and the glycohemoglobin (HgbA1C). Both the American Diabetes Association and the World Health Organization recommend confirming one test with another. The two-hour glucose tolerance test is generally not recommended except for pregnant women because its reproducibility is poor and its variance is large.
Once diabetes is diagnosed, stringent control of the glucose helps prevent the development of diabetic retinopathy (eye disease). Both the ADA and the WHO agree that a fasting glucose of 126 or greater or a glycohemoglobin of 6.5 or greater signifies diabetes. Both also characterize a fasting glucose of 100-125 as "pre-diabetes". The ADA calls a glycohemoglobin of 5.7-6.4 pre-diabetes, but the WHO does not, even though the test for HgbA1C is now standardized throughout the world.
The only clinically proven sustained method for overweight diabetics to lose weight and reverse their diabetes is by bariatric surgery---i.e. stomach banding or partial surgical bypass.
A retrospective clinical study published this week in JAMA showed that patients who are overweight when diagnosed with diabetes have a longer lifespan than patients who have normal weight when diagnosed.
There is no general agreement as to what age to start screening for diabetes or how often.
There are clinical conditions that can cause a falsely high or low HgbA1C. There also are medical conditions and drugs that can elevate or lower the fasting glucose.
Since diabetes is synergistic with other causes of coronary artery disease, many physicians would screen patients who have other risk factor(s) for heart disease: elevated cholesterol, high blood pressure, etc.
Blood pressure, cholesterol and fasting glucose often increase with age, and no one has yet shown if treatment of pre-diabetes really prevents diabetes (but the diagnosis may raise your life insurance premium).
Eating sugar will not make you diabetic unless you also are gaining weight.