Insulin treatment for type 1 diabetes mellitus (T1D) has improved over the past decades due to advances in insulin formulations and administration techniques. Nevertheless, optimal glucose control and reaching HbA1c goals remain challenging for many patients. Consequently, the risk of subsequent micro- and macrovascular complications remains high and life expectancy is still reduced in this patient group so that there is an obvious need for improvement of current therapies or the development of new, additional treatments [1]. So what else can be done?
A large body of current research focuses on potential adjunct therapies in type 1 diabetes, investigating a range of agents including SGLT2 inhibitors, GLP-1 receptor agonists, amylin analogues and others. To date, only pramlintide (Symlin®, AstraZeneca), an analogue of the pancreatic hormone amylin, has made it to the market, that is the US market only, in 2005. But which potential has this substance for therapy improvement in this patient group?
In contrast to healthy individuals, patients with T1D do not exhibit a significant amylin secretion or are even amylin deficient due to autoimmune destruction of β-cells. Natural amylin is produced by the β-cells and co-secreted with insulin in a fixed ratio in response to caloric intake. Interestingly, obese and non-obese individuals with impaired glucose tolerance also exhibit altered amylin secretion in a similar fashion as already recognised for insulin secretion (e.g. delayed and increased amylin secretion) [2].
This deficiency as well as the natural effects of amylin such as suppression of glucagon secretion, slowing of gastric emptying and regulating satiety suggests that amylin replacement therapy may have benefits in T1D therapy. Since amylin in its natural form is insoluble and tends to aggregate, the analogue pramlintide has been developed.
Clinical trials demonstrated that adjunct pramlintide to insulin therapy can decrease postprandial glucose excursions and improve HbA1c without weight gain [3].
Postprandial hyperglycemia was shown to be significantly reduced with pramlintide treatment in several trials. The time to maximum glucose concentration was prolonged, so that in combination with insulin pramlintide treatment can increase the risk of postprandial hypoglycemia. In early trials, severe hypoglycemia occurred after treatment initiation during the first 4-6 weeks [4, 5]. The effect of pramlintide on insulin doses has been studied in several trials suggesting that reduction of meal related bolus doses rather than basal insulin therapy needs to be adjusted [6, 7]. In order to prevent severe hypoglycemia, the manufacturer recommends to reduce mealtime insulin doses up to 50% at treatment initiation.
Positive effects of pramlintide adjunct therapy on HbA1c was observed in randomised clinical trials with a treatment period up to 52 weeks with a difference of approx. -0.3% at treatment end compared to placebo [8, 9]. In addition, the trial by Whitehouse et al included an open-label extension period of one year, during which the HbA1c reduction could be sustained, suggesting that the effect on HbA1c is not only a temporary one. Noteworthy, three times more participants in the pramlintide group of this trial reached an HbA1c level of less than 7% compared to the placebo group.
However, the effect of pramlintide on HbA1c is not consistent throughout trials. The trial reported by Edelman et al did not observe a difference in HbA1c after 29 weeks of either pramlintide or placebo treatment (approx. -0.5% in both groups) [6]. One difference of this trial compared to previous ones was that mealtime insulin doses were initially reduced by 30-50% in the pramlintide group. At the end of the treatment period, a reduction in total daily insulin doses (-12% pramlintide vs. +1% placebo) together with lower postprandial glucose values compared to placebo was observed.
The effect on body weight was comparable in the above-mentioned trials and patients lost weight in the pramlintide groups despite reductions in total daily insulin doses, while patients in the placebo groups gained weight. This weight loss is independent of nausea, a common transient site effect during the first 4-8 weeks of treatment. Ratner et al stratified the patients in their trial based on BMI and could show that the higher BMI group lost more weight compared to patients with a lower BMI [8]. This suggests that pramlintide may have benefits particularly for overweight T1D patients.
Pramlintide is not widely used and as mentioned before only available on the US market. From early postmarketing data a high discontinuation rate within the first 2 months of treatment initiation was indicated as patients seemed to be disappointed about treatment effects (e.g. insufficient weight reduction, side effects, lack of effectiveness, complicated dosing) [10].
Due to its pharmaceutical properties, pramlintide cannot be mixed with other short-acting insulin formulations thus requiring an additional injection prior to meals. Moreover, in contrast to the physiologic co-release with insulin in a fixed ratio, dosing guidelines for pramlintide recommend a fixed dose of 30 or 60 µg in addition to varying insulin amounts, an approach that leads to an inconsistent dose ratio compared to physiological conditions.
The group around Riddle et al. tried to mimick near physiological conditions and administered a fixed-ratio of pramlintide and insulin (via two wearable pumps with subcutaneous delivery) to 34 subjects with T1D over 24 hours. A fixed ratio of 9µg pramlintide per IU of human insulin was selected based on a published computer simulation that calculated the optimal ratio for a co-formulation [4, 5]. This fixed ratio was compared to a placebo and insulin infusion. CGM measurements were collected in addition to laboratory measurements of BG over the 24-hour periods. Co-administration of pramlintide in the fixed dose-ratio resulted in blunted BG excursions after all three meals (i.e. breakfast, lunch and dinner) in comparison with placebo. This was seen in both the CGM profiles and the laboratory blood glucose measurements. Moreover, pramlintide compared to placebo treatmentsignificantly reduced time in hyperglycemic range (≥180 mg/dL, 31 vs. 46%) and increased significantly time in target range (>70 to <180 mg/dL, 62 vs. 50%) while time in hypoglycemic range was not significantly different between treatments (≤70 mg/dL, 7 vs. 4%). As it could be expected, subjects during pramlintide treatment suffered more from gastrointestinal adverse events than during placebo treatment (47% vs. 7%). Overall, the authors conclude that this approach of a fixed-dose administration markedly reduced post-prandial glucose excursions as well as glycemic variability.
Although these results seem to be promising, it has to be kept in mind that this trial had a short duration and was performed under well-controlled conditions in a clinical setting with standardised meals. The potential of a fixed dose-ratio under real life conditions with varying meal portions and compositions needs to be evaluated thoroughly in future studies, in particular in the light of the potential hypoglycemia risk with adjunctive pramlintide treatment, and its tolerability. Moreover, a co-formulation of insulin and pramlintide seems a requirement to advance this adjunctive treatment further for pump treatment. The use of two separate CSII systems is first not cost-effective and second would presumably not be accepted by patients.
In conclusion, current evidence of the effect of pramlintide as adjunct therapy for the treatment of T1D hints to a beneficial role of pramlintide, but future work is required to establish its role in routine clinical practice.