By the time Type 2 diabetes is diagnosed the health-care situation can be considered 'late', as many people already have vascular damage, and in the remainder only by managing to target from diagnosis ('start early, continue long') will the benefits of glucose-lowering in reducing further vascular risk be obtained.  Partly this is a function of the so-called legacy effect, an epidemiological phenomenon reflecting that the early enabling vascular (endothelial) damage resulting from hyperglycaemia is reflected on a background of increasing adverse outcomes with age. 

That observation results in a challenge to late management, in the sense of management of hyperglycaemia many years after diagnosis.  Firstly there is a stable door problem – unless glucose control has been good in the prior decade, reversing the resulting damage may take a further 10 years.  Secondly the passage of time means that for some individuals the time of opportunity is lost, their life expectancy now, due to diabetes related vascular damage or other concomitant disease, being too short for the longer term gains to be attained.  Furthermore the underlying metabolic problems generally worsen with time, in particular insulin secretory capacity, so that the therapy load required to achieve metabolic targets will now be much higher.  A last issue here is that, if therapeutic targets and lifestyle changes have been suboptimal since diagnosis, it will feel perverse, and a mark of failure, to the person with diabetes to have sudden changes imposed. 

This leads to two issues in clinical practice, for both of which individualization according to need is important.  The first relates to agreeing glucose control targets.  We know in particular from the ACCORD study that tight blood glucose control is strongly associated with better mortality outcomes, though also from UKPDS that the benefits take 8-10 years to demonstrate.  ACCORD also tells us that people whose control is not easy to improve, people with more advanced complications, and people with more hypoglycaemia (paradoxically associated with poorer control as measured by HbA_1c) do badly.  These factors are a clear pointer to softening of glucose targets in those with shorter life expectancy and other medical conditions, though personal preferences over therapy burden are also important considerations. 

Issues associated with insulin insensitivity tend not to change with time from diagnosis, and accordingly fundamental management of that by lifestyle means is unlikely to be any more successful years after diagnosis than at the time of diagnosis.  Unfortunately the evidence does suggest that diabetes management everywhere is presently so poor that there still can be room for further lifestyle input late on, though if it has been previously optimized then delaying further therapy on the basis of some hoped for promise of change is probably bad practice. 

Present medication approaches offer a wealth of alternatives, though if islet B-cell function has progressed far, as shown by a history of deteriorating control (as opposed to continued poor control) despite appropriate use of oral therapies, then insulin injections offer a safe and easy option.  Indeed one approach to simple blood glucose management in the longer term is earlier use of insulin.  Otherwise choice of therapies will depend on such factors as a need to avoid hypoglycaemia, importance of body weight loss, and difficulties with tolerating medications.  Combining approaches, such as with a GLP-1 receptor agonist and insulin, or a thiazolidinedione with an insulin secretagogue/incretin/insulin can be very effective, but resource issues and in some circumstances safety/tolerability issues are real barriers.