Abstract
It is first fair to ask why in this era of rapid-acting insulin analogs and “smart” insulin pumps this type of advance even would be required. The Helmsley Trust T1D Exchange can give us some answers. Severe hypoglycemia is still too common in type 1 diabetes. The lowest rates are in the 6–12 year olds, at about 10 episodes/100 patient-years, whereas the highest rates are in those with long durations of diabetes over the age of 50 years; for this group, the rate is 50 episodes/100 patient-years. 4 By way of comparison, the event rate in the Diabetes Control and Complications Trial was 62 events/100 patient-years (in those patients receiving multiple daily injections or CSII). 5 So for this latter group of patients with the most impaired counterregulatory responses to hypoglycemia, 6 rates of severe hypoglycemia are unacceptably high and not different from a younger population of type 1 diabetes patients 20 years ago. More to the point, I would argue, any episode of severe hypoglycemia is unacceptable.
So the question still remains, is it possible to better integrate pump and sensor technology to reduce or, ideally, eliminate severe hypoglycemia? In this issue of Diabetes Technology & Therapeutics, Garg et al. 7 report a randomized crossover trial using a Medtronic (Northridge, CA) sensor-augmented insulin pump system with a low glucose suspend (LGS) feature that automatically stops insulin delivery for 2 h following hypoglycemia (<70 mg/dL). In this study, called ASPIRE (Automation to Simulate Pancreatic Insulin REsponse), hypoglycemia was induced with exercise with the LGS feature either turned on or off. For the 98 sessions where hypoglycemia was induced, hypoglycemic exposure was significantly reduced with the LGS system turned on (a reduction of 32.2 min).
Because this was a randomized crossover trial, it is not surprising that there were no differences in duration of hypoglycemia for the post-crossover sessions. Central nervous system adaptation to hypoglycemia occurs quickly. 6 Furthermore, “rebound hyperglycemia” was not observed with the insulin infusion suspension. This should also not be surprising given the often-unappreciated long duration of our “rapid-acting” analogs 8 and previous data showing lack of metabolic deterioration with insulin interruption at 0300 h. 9 In this older study with even longer insulin lispro interruption, after 2 h there was minimum impact on glucose or β-hydroxybuyrate levels. 9 This is important because the most critical time for an LGS system would be in the middle of the night. Outside of the research setting, insulin interruption (both intentional, such as what happens with longer than anticipated exercise sessions, and unintentional, such as catheter dislodging) is a common occurrence. In our practices we see insulin interruption greater than 2 h on a daily basis, almost always without significant consequence.
Why is this study so important? First, it confirms, in a well-controlled trial with hypoglycemia induction from exercise, that our current technology using LGS is adequate to reduce the duration of hypoglycemia in adult patients with type 1 diabetes already using CSII. Moreover, it seems reasonable to assume that as CGM technology improves, even more hypoglycemia will be prevented. Second, although hypoglycemia was minimized in this study, it was not prevented. Given the massive peripheral glucose disposal with exercise and the long duration of our current insulin formulations, this was no surprise. Finally, and perhaps most importantly, this current system was shown to be safe. This is also expected given the kinetics of the insulin during a 2-h suspension of insulin as noted above.
What next? Many of us are still in scratching our heads that this system using LGS has been available in the United Kingdom since September 2009 yet is still not approved in the United States. Even though the frequency of severe hypoglycemia has decreased with improved insulin delivery strategies and the introduction of insulin analogs, hypoglycemia remains far too common. In the secondary cohort (baseline glycosylated hemoglobin <7%) of the Juvenile Diabetes Research Foundation CGM study, hypoglycemia (<70 mg/dL) was present 6.3% of the time (1.5 h/day). 10 In another population of well-controlled type 1 diabetes patients, nocturnal hypoglycemia occurred in over half of the patients. 11 What we do not know definitively is mortality from hypoglycemia in our current era of insulin therapy. But we know it occurs, and even one death is too many.
In June 2011, the U.S. Food and Drug Administration released a draft guidance for industry for investigational device exemptions and premarket approval applications for LGS systems. 12 Required patient populations for study, primary end points, and trial designs were defined. However, because this is a draft guidance, it is unclear what the pivotal trials will look like. There has been much concern from the diabetes community about these initial guidelines, but we still await a final draft. Meanwhile, money continues to be spent on new studies that may not adequately address relevant questions, patients who could benefit from this technology remain without access, and, more chillingly, the sponsor could find the business strategy so ominous that this entire area of research be abandoned. This worst-case scenario could portend a similar future for development of a true “artificial pancreas” should the regulatory pathways become too difficult and expensive.
In the meantime, the ASPIRE study provides one more piece of information that, indeed, LGS can reduce exercise-induced hypoglycemic exposure. It is hoped that there will be more data provided quickly so that those patients who can benefit from LGS will have access.