So far in this series we’ve looked at why incretin therapy might be a useful adjunct to traditional type 1 treatment, and also how the pharmacokinetics might be best used to manage type 1.
In this article, we look at the feedback from users of these treatments and some of the positives and negatives that have been reported.
Introduction
Following a brief review of available research on incretin-based therapies, (a meta-analysis from 2019 and a small study of 10 T1D patients) this appears to be the first investigation into use alongside Automated Insulin Delivery systems and currently the largest associated with current incretin-based treatments.
While this is a relatively small sample (n=17) we hope it will provide some indications of the effects people are seeing in use with Automated Insulin Delivery systems.
This is a retrospective study of the use of incretin-based therapies with open source automated delivery systems in the real world.
The approach taken was a brief survey published in Facebook groups associated with Open Source AID systems, and asked questions relating to system, incretin, Time in Range (TIR), Time in Tight Range (TITR), whether users include mealtime management, changes in Total Daily Dose and any other commentary that users wanted to provide.
There was also a question about any bad side effects that may have stopped users from continuing to use a treatment.
Amongst those that responded, the biggest impact was a reduction in total daily insulin dose, with 91% of those reporting TDD data seeing a mean reduction of 45%. 59% of respondents reported a reduction in carb announcing/bolusing around mealtimes. Additional benefits included an increase in both TIR and TITR and reduction in hunger.
Adverse side effects included elevated ketone levels, gastrointestinal issues, mental health problems and difficulty in recovering from hypoglycaemia.
Further details can be found by reading on.
Method
A basic survey asking the following questions was posted into the following Facebook groups:
- Loop and Learn
- AAPS Users
- Trio (formerly an iAPS group
The survey consisted of the following questions and responses were gathered after five days:
- Which sensitivity method is in use for each system
- Incretin type
- Are meals being bolused for?
- Time in Range
- Time in Tight Range
- Any changes in TIR (with numbers if available)
- Any changes in TDD (with numbers if available)
- Any adverse side effects
17 users responded to the survey, both within the groups and also via direct messaging.
Basic descriptive statistics were then generated using Excel.
No socio-economoc or demographic data was captured and the period of use of incretins was not asked, although multiple users volunteered this information.
Results
Population data
The responses came from users of Loop, AndroidAPS and Trio.
The vast majority (94%) were using an oref1 based system, which is not reflective of the general user population of Open Source AID systems.
The split between Ozempic and Mounjaro was reasonably even.
Ozempic had the slight advantage, but this may be due to the period of time some respondents have been using incretin-based therapies.
Dosing regimes showed varying amounts across those who responded. Given the difference in dosing amounts, we’ve split them into tiers, where tier 1 is the entry level and tier 5 is the maximum amount.
The median amongst users sat squarely in tier 2, with very few people using more than this. Most users were in either Tier 1 or Tier 2. Another point that stands out from the responses is that approximately 1/3 of those responding have elected to titrate their doses to twice per week rather than once per week.
Total Daily Dose
The majority of users reported a significant drop in total daily dose, with a mean reduction of 45% (range 30%-55%).
Mealtime bolusing
Many users reported changes in how they manage mealtimes. The graph below shows the split of how this has changed.
59% of respondents reported that they had either stopped bolusing completely or that they had reduced the amount of bolusing required. Additionally, commentary suggested that for certain high carb or large meals, some users preferred to manually help the AID system along.
Time in Range and Time in Tight Range
Whilst all participants provided time in range (TIR) and time in tight range (TITR) data, few were able to provide pre- and post-use data.
Mean TIR was 89.3% (range 78%-98%).
Mean TITR was 72.6% (range 60%-86%).
Where pre-incretin use data was provided, the mean change of TIR was 11.9% (range 5.5%-26.7%), although the median change at 7.8% is perhaps more reflective of the results.
The mean change in TITR was 7% (range 6.1%-11%).
Qualitative factors
A number of items were highlighted by participants as things they considered benefits of using incretin-based therapies.
These include:
- Reduced hunger, both when in range and also during hypos
- Significant weight loss
- The ability to leave the system to manage meals
Adverse side effects
A number of adverse effects were identified by users:
- Elevated ketone levels – especially following fasted periods, where ketones were measured at levels associated with nutritional ketosis at normal blood glucose levels
- Gastrointestinal issues including both gastrointestinal pain and debilitating acid reflux
- Mental health issues, including anxiety
- Challenges recovering from hypos, with reaction to treatment taking longer
Discussion
A key aspect of this investigation is the cohort of respondents. All of them are selected from the Open Source AID, WeAreNotWaiting, community. Bearing this in mind, the level of engagement of these participants compared to the general population will be relatively high. This is reflected in the pre-treatment TIR provided, which was generally in mid-to-high 80 percents.
As a result, whether the changes in TIR and TITR fairly reflect what’s likely to be seen in a broader population requires further investigation.
What was also noteworthy was the dose sizes of incretins that were in use. Very few users had progressed beyond the first or second stage treatment levels (with some users having been on incretins for years).
This, combined with the titration that some users had undertaken, suggests that optimal dose size for managing type 1 is not aligned with the dose sizes designed for weight loss. There was also feedback from a number of users that they’d lost weight very rapidly at these low doses, which raises questions about type 1 diabetes and super responders.
Total Daily Dose
What’s clear is that regardless of the incretin in use, the total daily dose of insulin has dropped significantly in nearly all cases.
But does this matter?
Whilst it suggests a reduction in insulin resistance and metabolic syndrome related symptoms, it may also provide parallel benefits relating to the effects of insulin on board (IOB). There was mention of exercise being easier as a result of lower IOB.
It is possible that this is linked to the weight loss and reduction in food consumption that users experience, and is not the therapy itself generating this outcome, however that would require further investigation.
Interaction with the AID system
There are three items of interest in the interaction with AID systems.
The first is that the vast majority of respondents used either AndroidAPS or Trio. There were very few responses from the Loop community.
It’s possible that we simply didn’t reach the Loop community in the same way that we reached the other two. It may also be the case that the type of user attracted to the two systems highlighted are more likely to experiment with adjunct therapies.
The other possibility is that the unannounced meal capabilities in the oref1 based systems attracted users to incretin based therapy with the idea that they may be able to manage levels more easily with less interaction.
Anecdotal feedback also suggested that moving on to these therapies using Loop was harder than oref1 based systems because of the way that Loop handles meals.
This brings us on to the second point.
59% of users stipulated that they had reduced mealtime interaction when going on to incretin-based therapies. 18% stated that they had moved to a “no bolus” approach. This suggests that the potential benefits that were highlighted in the first article of this series have been brought into play.
Whether this is aided by the unannounced meal functionality within oref1 as it exists in AAPS and Trio is unclear, however, it suggests that users benefit from having less to worry about around meals with the use of incretins alongside AID systems, but raises the question about the requirement for meal detection.
Finally, we already noted that a minority of users have started titrating to a split dose, to soften the variation from a single dose. Open source AID systems are all, at their core, fairly simple and don’t contain artificial intelligence or machine learning.
If the traditional single dose model was used with systems that do have a temporal learning function, how might this affect its behaviour? Would it be better to start with split doses?
Adverse side effects
Aside from gastrointestinal side effects, which were mentioned, and are well known with these treatments, some other items are worth reviewing. It was notable that none of the respondents to the survey mentioned issues with gastropareisis.
The reports of higher than normal levels of ketones after prolonged periods with no eating raises questions.
- Are these seen in all users and need consideration in relation to periods where insulin levels might fall significantly, in order to avoid euglycemic DKA?
- Is it related to the rate at which weight is lost?
- Once in a steady state are the same ketone levels seen?
It may be that this is a side effect that temporarily occurs during weight loss as the body burns a high level of fat. If this is the case, then it should cease once a steady weight is achieved.
On the other hand, it may be an ongoing concern, at which point, the interaction with AID becomes important. We already see issues with existing AID systems during prolonged exercise where long, zero, temporary basal rates may be used. Could incretin-based therapies enhance these reactions? it appears to require additional investigation.
The other point of note within type 1 use is the increase in hypo recovery times that was noted. If a user does find themselves low, there are two potential factors coming in to play.
- Reduction in the alpha cells ability to trigger glucagon release to counter hypoglycaemia
- A reduction in the speed at which food, and especially fast carbs, can be absorbed through the gut
Given that the first of these is normally physiologically triggered at 3.6-3.9mmol/l (65-70mg/dl), it would make sense that the limiting of glucagon release might have an impact.
Combining this first issue with a slower release of fast carbs from an intervention would explain the phenomenon described. It suggests that alternative hypo treatments may be required.
It also raises the question of sensitivity to glucagon. Does the additional effect of blocking glucagon secretion enhance the liver’s response to exogenous insulin?, Could this allow microdosing of nasal glucagon, for example, to recover quickly from hypoglycaemia?
Takeaways
The messages that come out of this small study are:
- Incretin-based therapies work well alongside AID systems
- Initial feedback suggests that people do see significant benefits
- AID systems with some form of meal detection allow users to take a step back from self-management
- There is a lot we don’t know, and a lot that needs further investigation
- There are potential issues hiding in plain site that need to be further investigated
The key takeaway is that while this remains a medication that is not indicated for type 1 diabetes, a lot of these factors remain uncaptured.
We need, at the very least, the ability to capture and reflect the data described here more easily. In addition, trials that follow up some of these questions would benefit all users.
Ultimately, this is a very new field with lots to learn and plenty of opportunity. It will be interesting to see where it leads.
Leave a Reply