The risk of developing complications in type 2 diabetes rises in line with increases in HbA1c – but when should insulin be started? Professor Andrew Farmer outlines the key considerations
The number of people with diabetes has doubled in 15 years. Cardiovascular disease is now the major cause of death and there is substantial morbidity from, for example, renal and ophthalmological complications.
In 1998, the results of the United Kingdom Prospective Diabetes Study (UKPDS) focused attention on the possibility that intensive glycaemic control for a period of 10 years, with therapies including insulin, might lead to reductions in myocardial infarctions for people with type 2 diabetes. More recent analysis after a further 10 years of follow-up confirms that, despite loss of differences in HbA1c between the trial groups, there is a 15% reduction in risk of MI and a 13% reduction in death rates with intensive treatment.1
Aggressive management of all cardiovascular risk factors, including hypertension and dyslipidaemia, remains the best way to reduce risk. The recent Action to Control Cardiovascular Risk in Diabetes (ACCORD) Trial has raised concerns about intensive and rapid control of glycaemia in patients with long-standing disease, but there is still a strong evidence base for starting insulin in patients who are not at HbA1c targets (below 7.5%).2,3
When to consider starting insulin
Insulin treatment is recommended as third-line therapy for diabetes following use of metformin as first-line and sulphonylureas as second-line.
Other situations in which insulin should be considered include pregnancy or where there is a strong patient preference.
Working closely with a diabetes specialist nurse or a practice nurse experienced in insulin initiation is invaluable in ensuring good support for the patient.
Insulin treatment may become inevitable, even among those whose concerns about insulin have meant they have decided not to start it in the past. Waiting until there is further deterioration in control and consequent symptoms of polyuria, polydypsia and oral thrush, with associated weight loss, is associated with further complications.
Although many patients will benefit from improved glycaemic control with insulin, for some insulin therapy may be unacceptable or inappropriate. Patients for whom insulin may be inappropriate include those with:
• a BMI of 35kg/m2 or more (in those of European descent with adjustment for other ethnic groups)
• specific psychological or medical problems associated with high body weight
• a BMI below 35kg/m2 for whom insulin treatment would have major occupational implications, such as lorry drivers.
Important considerations before starting insulin treatment are outlined below. In particular, the added burden of hypoglycaemia, which might be a particular problem in elderly or frail people, may mean an alternative should be tried.
Choice of insulin
Practice varies widely, with differing regimens recommended for initiating insulin therapy.
When insulin treatment is added to oral therapy as glycaemic control becomes inadequate, a long-acting human insulin is appropriate and should be injected at bedtime or twice daily as necessary.
An analogue insulin – such as insulin determir or insulin glargine – may be appropriate if the person needs help in injecting, and potentially avoids the need for two injections.
Patients starting on a basal insulin may eventually need to move to a basal-bolus regimen to attain overall glycaemic control, replacing their sulphonylurea with a short-acting insulin.
The recent 4-T trial compared patients with type 2 diabetes in whom treatment was started with either a long-acting insulin (determir), a biphasic analogue insulin or a pre-meal regimen. Patients started on the long-acting insulin had better outcomes in terms of glycaemic control, lower weight gain and fewer hypoglycaemic episodes after three years.4
Calculating an initial dose
Conventional insulin regimens have been based on starting with 10 units and then titrating upwards. But as many patients will need much higher daily insulin doses than those used in type 1 diabetes, and this can lead to a long period of titration.
The 4-T trial used an algorithm to predict likely insulin dose and suggest a starting dose closer to the likely required dose, with some patients starting on over 50 units of insulin.4 However, until algorithms for the starting insulin dose are better defined, it is better to start at a low dose and work upwards, explaining carefully that a higher dose may be required to avoid concerns or disappointment with increasing dose. Dose increases are usually suggested in 10% increments (see table 2).
Agreeing a management plan
• Following a structured education session, telephone contact needs to be established and maintained to allow discussion of required insulin dose changes.
• If not already self-monitoring blood glucose, then the techniques required need to be taught and a system for recording the results established.
• Arrangements for disposing of sharps need to be put in place.
• Administrative issues, including the need to inform the DVLA, should be discussed.
• The patient’s understanding of diet needs to be explored. Weight gain is common for several reasons including lower urinary losses of glucose with better control.
• Information also needs to be given about the possibility of hypoglycaemia and the need to test blood glucose when driving, and have available a plan for dealing with hypoglycaemic episodes.
• Special considerations must be given to those with a need to fast (for example during Ramadan) or those who do shift work.
Professor Andrew Farmer is professor of general practice at the department of primary health care, University of Oxford
Competing interests: none declared
Table 1 Table 2 Table 3 Insulin initiation