Original article:
Pharmacokinetic and pharmacodynamic properties of long-acting insulin analogue NN304 in comparison to NPH insulin in humans. Brunner GA, Sendlhofer G, Wutte A et al. Exp Clin Endocrinol Diabetes 2000; 108: 100–5.

Summary
NN304 (insulin detemir) is a long-acting insulin analogue. It is formed by removal of threonine at position B30, and the addition of a C14 fatty acid at B29. NN304 is 98% reversibly bound to free fatty acid binding sites on albumin in plasma and interstitial fluid. This unique mechanism of albumin binding prolongs its duration of action, and contrasts with existing long-acting insulins whose duration of action is dependent on the rate of dissociation of various sized crystals at the subcutaneous site. The biologically active insulin is the free fraction of NN304 that binds to the insulin receptor.
NN304 is also unique in that it is soluble, i.e. a clear solution, and thus avoids the problems of mixing with other soluble insulins and of inadequate suspension (prior to injection) that can occur with existing crytallized (cloudy) long-acting insulins.
This study applied the glucose clamp technique to 10 non-diabetic subjects. The effect of 0.3 and 0.6 U/kg of both NN304 and NPH were compared.
The levels of serum insulin were similar, but this included the bound fraction, as it is not currently possible to measure just the free insulin. The action of the insulin was therefore assessed mainly by the rate of glucose infusion required to keep the blood glucose level constant (Fig. 1). As can be seen from the figure, the results appear to confirm that the action of NN304 shows less of an early peak than NPH. As can also be seen from the figure, the effect of NN304 was about 30% that of NPH. There was also much less of a dose-response effect with NN304, i.e. doubling the dose did not have much effect, in contrast to NPH.

Comment
Although fast-acting insulin analogues have improved daytime control, and even reduced the rate of nocturnal hypoglycaemia, the remaining problem is often control of the fasting blood glucose. This depends on the level of circulating basal insulin at that time. At present, this basal insulin is usually provided by a dose of NPH delivered in the evening. Ultralente insulin is sometimes used as an alternative, in the hope that it will work for a little longer. However, in contrast to insulin of animal origin, the biosynthetic version of ultralente insulin differs only slightly from NPH. One study confirmed the general impression that NPH provided better control than ultralente.
The clinical challenge is to find a dose of evening insulin sufficient to control the fasting blood glucose without causing nocturnal hypoglycaemia. As can be seen from the figure, NPH insulin has a peak action of around 4–6 h after injection, leaving insufficient insulin to act before breakfast. The difficulty is compounded by the well-documented dawn phenomenon (‘as the sun rises so does the blood glucose’), due to the increased secretion of hormones such as cortisol and growth hormone at that time. More insulin is thus needed in the hours before waking than is required during the earlier part of the night. Strategies such as changing the timing and site of insulin administration may be used to counter this problem, but the results vary in individual patients.
The mechanism of an insulin analogue binding to albumin provides a depot of insulin in the blood rather than the subcutaneous tissue. Early studies of NN304 such as this one hold out the possibility that such a mechanism may provide more insulin when needed, i.e. before waking, rather than peaking earlier in the night, as with NPH.
Programmable insulin pumps (continuous subcutaneous insulin infusion, CSII) are effective because the basal rate can be programmed in advance to increase (or decrease) when needed. New long-acting insulins such as insulin glargine and NN304 will probably never be able to match this degree of basal insulin adjustment, but they may do a better job than existing insulin preparations.
Another problem with currently available long-acting insulins is the variability in absorption rate at the subcutaneous site, resulting in large day-to-day variations in individuals’ fasting blood glucose levels. Recent studies suggest that there is much less of this variability with NN304.
Other studies have confirmed that the effect of NN304 is about 25–30% that of NPH, and this appears to be due to reduced affinity to the insulin receptor. This should not be a problem in clinical practice where the dose can be adjusted as required. However, it might be helpful if the concentration (in international units/ml) of NN304 were calculated on the basis of biological activity (i.e. a return to the rabbit method) rather than the measured amount.
Whether the lack of dose-response seen with NN304 is due to the lower effective dose used or some other mechanism remains to be seen.
Further studies are underway to clarify the usefulness of NN304 as a basal insulin. Hopefully, we will, in the near future, have a new range of long-acting insulin analogues to add to our existing therapeutic armamentarium of fast-acting insulin analogues that have already proved their usefulness.

Summary and Comment:
Matthew Cohen, Caulfield, Victoria, Australia