Appendix B
Derivation of Appropriate
Dual Ratio Amylin/Insulin Dosing
Since the purpose of hormone replacement therapy is to mimic healthy endogenous plasma levels, it follows that subcutaneous dosing of hormone agonists should be designed to achieve these differences in plasma basal and bolus ratios. Since T1D patients with diabetes adjust their daily insulin dosing to reflect exercise and eating patterns, the best way to accomplish this is to determine amylin agonist doses based on amylin’s healthy physiologic relationship to insulin. In this way patients would have a simple algorithm for deciding how much pramlintide to take based upon their individual insulin doses.
In this appendix we use the following process to estimate ratios for the Dual Ratio Amylin/Insulin (DRAI) dosing:
Estimate exogenous dosing ratios to mimic plasma profiles: Based upon the mass ratios of insulin and amylin plasma levels, we first estimate the amounts of amylin agonist that should be infused subcutaneously calculated as ratios to the basal and bolus insulin doses.
Convert mass ratios to convenient guidelines for mixing with insulin: We then convert the mass ratios to µg/U ratios and illustrate how these would dictate pramlintide dosing as a function of insulin dosing.
Estimate Exogenous Dosing
Ratios to Mimic Plasma Profiles
Insulin-dependent dosing can be achieved by calculating two separate components of pramlintide dosing: (1) the weight ratio of the daily basal component of amylin-to-insulin; and (2) the weight ratio of the daily bolus component of amylin-to-insulin. Table 1 shows this calculation: column two shows the calculated molar AUCs for endogenous plasma insulin and amylin; column four shows endogenous gram AUCs after correcting for molecular weight differences; and column six shows exogenous gram AUCs after correcting for bioavailability of subcutaneous (SC) injections. The exogenous ratios of pramlintide-to-insulin are highlighted in grey.
These calculations indicate that, to mimic healthy endogenous hormone levels, basal SC infusions of insulin should be accompanied by about 15% pramlintide by weight to insulin, and bolus SC injections of insulin should be accompanied by about 5% pramlintide by weight to insulin.
In the example above absorption and clearance rates for insulin, amylin, and pramlintide are assumed to be approximately equal. For use in a dual hormone pump, the most likely choice of insulin would be from among the rapid acting varieties, e.g. APIDRA with an apparent SC half-life of 42 minutes compared to the apparent SC half-life of 48 minutes for SYMLIN; in this case the assumption of about equal absorption and clearance rates is probably appropriate. In practice, actual dosing ratios should be determined by taking into account the specific pharmacokinetics of the insulins and amylin agonists being used.
Convert Mass Ratios to Convenient
Guidelines for Mixing with Insulin
Because insulin is traditionally dosed in Units rather than µg, it is more useful to express these ratios as µg of pramlintide per Unit of insulin. In Table 2 the WHO standard for insulin is converted into µg per Unit of insulin, which is then multiplied times the basal and bolus ratios highlighted in grey in Table 1.
Thus we have the DRAI ratios:
Basal: 5.7 µg pramlintide per Unit basal insulin.
Bolus: 1.8 µg pramlintide per Unit bolus insulin.
A Real Life Example of
Dual Ratio Amylin/Insulin Dosing
To put some flesh and blood around the above analysis, following is the experience of a 50-year-old man with T1D who has tried and abandoned pramlintide. This individual is well above average in education and medical skills: he is an MD who is presently CEO of a biotech firm in the molecular biology area and who has remained on the faculty of a top tier university. He has collaborated in this analysis of pramlintide dosing from the perspective of someone who recognizes the theoretical benefits but found the cost/benefit ratio to be unacceptable.
He started with 48 µg mealtime doses of pramlintide while backing off insulin from 40-50 U/day to 30-40 U/day, a reduction of 20-25%. Because of nausea he reduced pramlintide to 24 µg doses, or 72 µg daily (this was before vials were replaced by prefilled syringes with set amounts). During the nine months he was on pramlintide he experienced significant though short-lived (30-45 minutes) nausea even at the lower doses, although he never felt close to vomiting. Because he had already been maintaining tight blood sugar control (HbA1c at 5.7-6.4), he was concerned about the time it took to pull out of hypoglycemic episodes. Over the nine month period he lost 15-20 pounds, so discontinuing was a tough call. In spite of his recognition of the theoretical benefits of pramlintide therapy, its burdens simply did not justify continuing because he could achieve his HbA1c target with insulin alone.
How would he dose pramlintide using an AID system and the DRAI algorithm? Let’s run the numbers…
At the upper range of our subject’s daily insulin dose with 50% as bolus, his average basal/bolus doses would be as shown in Table 3.
The indicated daily dose of pramlintide would be 150 µg:
Basal dose: 114 µg would be infused as a continuous basal level over 24 hours, i.e. at an infusion rate of about 4.8 µg per hour.
Bolus doses: 36 µg would be infused with the three insulin boluses, i.e. 12 µg at mealtimes, assuming equal mealtime insulin boluses.
Following package insert instructions, he was taking mealtime injection boluses of 24 µg for a daily total of 72 µg. In other words:
Mealtime overdosing: During a 2-hour post prandial period he should have received [12 + (2 x 4.8)] = 21.6 µg, ideally with the 12 µg as a square wave bolus to better match the natural amylin profile. Thus his mealtime injections of 24 µg were about 10% over this starting ratio; either his continuing nausea was explained by this overdosing, or his ideal bolus ratio to avoid nausea might be lower than 1.8 µg/U.
Daily underdosing: During a 24-hour period he should have received 150 µg, or twice the amount that resulted from following the package insert instructions. This would explain the disappointing impact on HbA1c, and the lack of any noticeable impact on his hypoglycemia experience.
The doses calculated from the in vivo plasma profiles are different from current clinical practice. However, the most dramatic difference is not dosing size, but rather dosing profile: it is changed so that only about one quarter to one third of the daily dose is administered as mealtime boluses, rather than 100% of the daily dose. By emphasizing the basal component of amylin’s plasma profile, efficacy should be maximized without triggering nausea, and the more natural plasma profile will reduce insulin dosing, which could be expected to correct the problem of hypoglycemia stickiness caused by too much insulin onboard.
In actual clinical practice, the DRAI ratios would be adjusted to (1) avoid any mealtime nausea while suppressing postprandial glucagon and (2) achieve a high enough basal level to restore glucagon counterregulation.