In the only other single-center study in adults with T1DM (mean age 45 years), Faulds and colleagues evaluated changes in glycemic control and adherence in patient using the same hybrid closed-loop system.14 Treatment resulted in a decrease in HbA1c compared with baseline similar to our study, most notably for patients who had higher baseline HbA1c. However, over its short duration (6 to 12 weeks), there was decreased time in automated mode in study patients, likely due to treatment burden. Our study in older patients showed a similar reduction in HbA1c from baseline up to the 7-month visit but with increased sensor wear and time in automated mode.
There are many possible reasons for improved time in target range in our older population. Contrary to common belief that older age may be a barrier to adopting complex technology, it is likely that older age and longer duration of DM motivates adherence to a therapy that reduces glucose swings, offers a greater sense of safety and control, and improves quality of life. This is underscored by improvements over time in sensor wear and time in automated mode, measures of adherence, and successful AIP management. In support of a motivation factor to adopt insulin pump therapy in patients with long-standing T1DM, Faulds and colleagues found that older age and higher baseline HbA1c were associated with less time spent in hypoglycemia.14
The close supervision of patients by a certified diabetes educator and pump trainer may have helped improve glycemic control. Veterans received initial training, weekly follow-ups for 4 to 5 visits, and then bimonthly visits. There was also good access to the DM care team through a secure VA messaging system. This allowed for prompt troubleshooting and gave veterans the support they needed for the successful technology adoption.
The use of real-time CGM led to improvements in hypoglycemia unawareness. The nature of automated insulin delivery not only allows the patient to use a immediate CGM, but automatically lowers the delivery of insulin, further minimizing the risk of hypoglycemia.15 This combined approach explains the improvement in self-reported hypoglycemia unawareness in our cohort which decreased by 61%. As in our study, very recently Pratley and colleagues reported in a 6-month follow-up study that the greatest benefit of CGM was not the -0.3% improvement of glycemic control (similar in magnitude to our study) but the 47% decrease in the primary outcome of CGM-measured time in hypoglycemia.16
Hybrid closed-loop insulin delivery improves glucose control while reducing the risk of hypoglycemia. There is consensus that this approach is cost-effective and saves resources in the management of these complex patients, so prone to severe microvascular complications and hypoglycemia.17,18 A recent analysis by Pease and colleagues concluded that the hybrid closed-loop system was safer and more cost-effective when compared with the current standard of care, comprising insulin injections and capillary glucose testing.19 This held true even after several sensitivity analyses were performed, including baseline glycemic control, treatment effects, technology costs, age, and time horizon. This is relevant to the VHA, which at all times must consider the most cost-effective approach. Therefore, while there is no such debate about the cost-effectiveness of AIP technology for younger adults with T1DM, this study closes the knowledge gap for middle-aged veterans.7,10,12,13 The current study demonstrates that even for older patients with long-standing T1DM, when proper access to supplies and support services are made available, treatment is associated with considerable success.
Finally, AIP is well suited for telehealth applications. Data can be uploaded remotely and sent to VA health care providers, which can facilitate care without the need to travel. Distance is often a barrier for access and optimal care of veterans. The current COVID-19 pandemic is another barrier to access that may persist in the near future and adds value to AIP management.
There were a few challenges with use of AIP. Although transition to AIP was smooth for most patients already on insulin pump therapy, several noted requests for calibration in the middle of the night in automated mode, which affected sleep. Also, AIP technology requires some computer literacy to navigate the menu and address sensor calibrations, which can be a challenge for some. Based on our results, we would recommend AIP in veterans who are appropriately trained in carbohydrate counting, understand the principles of insulin therapy, and are able to navigate a computer screen menu. Most T1DM patients already using insulin pump meet those recommendations, thus, they are good candidates.
Limitations
There are some limitations to our study. The small sample size and single-center nature prevent generalization. Also, the veteran population cannot be extrapolated to other populations. For instance, the majority of the patients in this study were male.
Conclusions
We report that an AIP approach for patients with long-standing T1DM is well accepted and engages patients into monitoring their blood sugars and achieving better glycemic control. This was achieved with minimal hypoglycemia in a population where often hypoglycemia unawareness makes DM care a challenge. Future studies within the VHA are needed to fully assess the long-term benefits and cost-effectiveness of this technology in veterans.