How closed loops and faster, liver directed insulins will change your life #wearenotwaiting

Imagine, if you will, that those who struggle with their diabetes need only minimal self-intervention to maintain DCTT recommended Hba1C levels and excellent glycaemic variability. They only have to strap a pump and sensor on and adhere to changing the cannula every three days and the sensor every fourteen. No calibration. No finger prick testing. No bolusing. They can forget about their diabetes. You think this isn’t possible? I think it could be here sooner than you think.

The current trials of Hepatic-Directed Vesicle insulin and the new faster-Aspart and BC-Lispro present an immensely interesting opportunity to remove the need to bolus for food as a Type 1 diabetic. With the appropriate algorithm, a Hybrid Closed Loop could become a straightforward closed loop!

Combining this with a calibration free sensor could provide a way for the most uninvolved, uninterested people with Type 1 Diabetes to achieve reduced glycaemic variability and Hba1Cs that their doctor could only dream of.

HDV and Faster insulins – the background…

Back in January, I wrote about HDV – an additive that allows insulin to be targeted at the Liver, where in a normal person, the majority of insulin is used. According to recent research, this is about 80% of all naturally produced insulin. In a Type 1, essentially no exogenous insulin is used by the liver as it is used by other cells throughout the body.

In an interview with Robert Geho, the CEO of Diasome, details of the current state became available. This product has now entered Phase 2b trials, which the company describes as a sub-set of Phase 3 trials, which have also been approved. The approvals have been given for a pre-mix with an insulin manufacturer and an “addition” material to go into existing insulin phials. Given the two options, they are undertaking phase 2b with the “add” mixture, rather than the pre-mix, which is designed to work with all forms of commercial insulin, which will give an indicator as to the direction that they plan to take for phase 3.

The phase 2b will focus on the human analogues pre-meal, and they will try to achieve a 0.5% drop in Hba1C levels. This is good news, as the technology can then be used in pumps as well. The effect of having the liver remove more glucose in theory means that the peripheral glucose levels will be lower and that less insulin will be needed. With less insulin, the expectation is that weight gain as a result of exogenous insulin will be reduced. Diasome is expecting to see a 25% to 30% reduction in insulin needs.  The previous testing showed a significant improvement in post-prandial glucose levels.


As a corollary, the reduction in insulin needs should result in lower hypoglycaemia. More interestingly, the dog trials that were done showed that using HDV insulin eliminated hypos due to better glucose storage in the liver and it being better available in the presence of glucagon.

Where this gets really interesting is the faster insulins that I wrote about mid-September. As you’ll recall, BC Lispro from Lilly showed some very promising results:


Let’s take a quick look at these two side-by-side. The BC-Lispro showed a massive reduction in glucose level excursion compared with standard Lispro, reducing the increase in level by about 50%.

The HDV-modified Humulin-R also reduced the level of blood glucose excursions, reducing the increase by around 35%.

So then, imagine what might be achieved by combining the two.

Entering the imaginarium

As readers will know, I’ve built an OpenAPS rig and also tried to turn the Abbott Libre sensor into a CGM. Efforts persist in the Open Source community on that front, but it can’t be too far off that Abbott integrate that calibration free technology into a new CGM product. The only other thing they need to do with it is bring the MARD below 10% to make it possible to get regulatory sign off for bolusing.

I’ve also been using the current dev branch of OpenAPS for a little over a week, with the Advanced Meal Assist (AMA) function playing an important part in the latest codebase. Why is it an interesting feature? Because you can forget to bolus for a meal and it makes a surprisingly good fist of dealing with it.

What is AMA doing? Well you tell it you will be eating soon by giving a small bolus roughly an hour ahead of eating, enough to bring your glucose level down to 4.0mmol/l, and then, if you forget to bolus, it starts to provide larger TBRs once it picks up your glucose levels rising. It does a surprisingly good job of this with the current insulins and manages to avoid massive spikes while also keeping you from plummeting to lows. It does need that trigger though.

So imagine a state where, instead of warning that you are about to eat, the increase in glucose levels is the trigger. And imagine that when this happens, the fast acting insulin that is released acts significantly more quickly, and on the liver. That’s where these two products are taking us.

By combining HDV with Faster Insulin, we may be at the point where those using a Hybrid Closed Loop with these sorts of advanced features won’t need to indicate that they are eating. Add in the idea of CGM with a calibration free sensor and you arrive at a  closed loop with minimal daily human intervention.

I dare to dream. For the 25% of people with Type 1 who really struggle? This can’t come soon enough. For the rest of us? We feel exactly the same.

Be the first to comment

Leave a Reply

Your email address will not be published.