Closed Loop for a Diabetic Cat???

Yes. For our cat.

Important – I am neither a veterinarian nor a physician, and nothing on this website or blog should be viewed as medical advice.

This is what I do for our cat Buddy.

I have opinions, and this being my blog, I’ll state those opinions. Whatever I write here, are my opinions, and is never medical advice. Your cat’s diabetes – and your opinions – may vary.

This blog is not sponsored by anyone. I name some specific products here and even link to their website – in this “space” of sticking an insulin pump and a CGM on a small animal with very thin and very loose skin – hoping that the very expensive devices don’t come off after a day – there are precious few products that work.

Aside from telling the story of Buddy the Cat, I think that my writing about what adhesive does not take my kitty’s skin off and holds his Omnipod in place for 3 days and his sensor for more than 10 days can inform others in a similar situation. 6 weeks ago, I had a lot of questions and did not find many answers. So, Buddy and I did some trialing and erroring, found some answers for us, and maybe that can help another >^..^<.

For the love of cats, if you consider doing this for your own kitty with diabetes – please inform yourself and work with your vet, inform yourself some more. I’ll put a page with resources together in the next few days – like Loop and Learn etc.

Important: Please don’t expect your vet to have knowledge about closed-loop systems, insulin pumps, and such. That’s neither their fault nor failing – insulin pump therapy simply does not exist yet as a standard treatment for cats with feline diabetes. Closed-loop insulin therapy is only starting to exist as an (approved) treatment for people.

I don’t think putting a cat on a “tubed” insulin pump like a Tandem or Medtronic pump is a good idea. But a tubeless pump such as an Omnipod is working just fine for Buddy (with the right adhesive and with diluted insulin). Insulet’s Omnipod 5 system might work for him, but I don’t have that system around.

From what I see, Omnipod 5 could be a good “commercial” system for people with diabetic cats who aren’t already involved in #wearenotwaiting and DIY Closed Loop stuff for themselves or their loved ones.

About me – I am a human with Type 1 Diabetes. I have been using closed-loop systems since 2018. Aside from one “hybrid closed loop”, none of the systems I have used are approved by the FDA. These are experimental do-it-yourself approaches to living better with diabetes. #wearenotwaiting. As a user of these systems, I owe deep gratitude to many people who have dared to develop, document, and keep improving on real solutions that the medical device industry has long not been able to provide. I’ll add links to the site to these resources.

When Buddy was diagnosed with Type 2 Diabetes on February 24, 2023, I put a Dexcom G6 on him the next day. I only waited that long because I was still trying to figure out how attach the very expensive CGM sensors to a cat in a way that it does not fall off within minutes, and I had to wait for the medical adhesive to be delivered (overnight). I needed to understand whether the sensor values were usable – these things are calibrated for human blood, after all.

Seeing his blood sugar behavior in real-time, I saw within hours that the treatment initially prescribed by his vet was not going to work for Buddy and for me. (Initially, he was on 2 units of Prozinc twice a day). The Prozinc worked on Buddy as Regular Insulin works on me – with a pronounced action curve and maybe 6 hours of duration of insulin activity. He went from 400 mg/dl down to around 80 within a couple of hours, at which time I gave him a small amount of honey to prevent the cat from getting a hypo. When the insulin stopped working way inside of 6 hours, he went right back to 400. To prevent him from going out of range of the CGM (400 mg/dl) I started augmenting with short-acting insulin (which I use in my insulin pump). Very stressful for Buddy and for me. I did not get many hours of uninterrupted sleep those first days. But, this way (the hard way …), we discovered how Buddy responds to the short-acting insulin FIASP and Humalog. (In short: Fine:! Fast onset of action, fast out, and extremely sensitive). That’s what you want in a closed loop system – the faster the onset and the shorter the duration of insulin activity (DIA), the better for a closed loop that approximates what happens in a non-diabetic body.

As an aside: On team T1D, we humans on closed loop systems are excited about the new ultrafast acting insulins such as Lyumjev and FIAsp, and want even faster onset and shorter DIA – so the CGM sees the blood sugar rise and can immediately and gently act on it – and if the BG starts dropping, there is not much insulin on board (IOB) to cause a hypoglycemia.

Anecdotally from looking at Buddy’s curve, it seems that he has advantages over me in both onset of action and DIA. Might be worth looking at more closely in a study.

His vet secondly prescribed Glargine (generic for Lantus), and I tried a couple of days of insulin Glargine twice a day with Buddy, but that did not work well (enough) either. Glargine is supposed to have an almost flat “curve” to its action, but that was not the case for Buddy. 2.5 IU of Glargine dropped him by 250 mg/dl, and the insulin was showing no more action 8 hours after injection, and his BG was running right back up toward 400.

Buddy was not in Ketoacidosis when diagnosed, but he clearly had not been feeling well before he went on insulin – he was throwing up, was listless, and depressed.

With the sensor, we now had BG data to add to the observations. I observed Buddy becoming lethargic and depressed whenever his BG was getting very high. I have first-hand experience with how very high blood sugar feels for me, and I imagine it looks similar to an observer – listless, depressed, and a bit confused. Not fun.

Buddy goes “Closed Loop”

With my own 4+ years of using OpenAPS, a dash of AndroidAPS and Loop, and in the last 2 years FreeAPS X and its new incarnation iAPS – the question was not “IF”, but “HOW” to use an Omnipod in a closed loop system on our cat with diabetes. Since I use iAPS, Buddy is using iAPS (I think it’s great – OpenAPS algorithms on an iPhone, ability to use Omnipods, plus plus. <3. More about this later.)

I had Omnipods, a Riley Link, and even Diluent for the short-acting insulin at home. His vet was on board, and prescribed Dexcoms and OmniPods. I called Buddy’s insurance, and they might actually cover the sensors and maybe even the pods. Cool if they do – we’re about to reach the deductible of $500 with the insulins and sensors. I have not picked up “his” Omnipods at the pharmacy yet – we’re still using some I had around. Those things are expensive, so I hope his insurance will cover them. Perhaps he’ll be in remission before we need more pods? (fingers crossed!!!). Anyways, Pod or Not – I expect to keep a sensor on him for a while yet.

Short answer to HOW:

• Appropriately diluted FAST-ACTING insulin in an Omnipod,
• Dexcom G6 sensor with an Anubis transmitter,
• iAPS to close the loop, and
• Skinister Medical Adhesive to keep the devices on the cat until they are ready to come off.

Also: Diet. He needs to lose weight (about 4 pounds!), and needs to eat a healthy low/no-carb diet.

• Restricted-calory low carb diet (aka species-appropriate – wet food, canned or raw. NOT dry food/kibbles).
• You can buy good to excellent wet food at a pet store, online, and at grocery stores. It comes in cans, little aluminum foil trays, and foil pouches.
• “Gravy” foods are much higher in carbs; we buy pâté (like Costco Pâté) and “chunks” without gravy – such as TikiCat After Dark). I sometimes make raw food, and have cooked some food for the cats.

The biggest question was at that point not a medical question, but one of ADHESION – How to keep an OmniPod securely on the cat for 3 days, and a sensor for at least 10 days. I tried VetBond – a tissue glue (You could call this “medical Superglue”, and, no – don’t use superglue …). VetBond works very well to glue the Dexcom sensor and Omnipods to the cat’s skin. I think, “too well”. It’s not ideal for long-term use, and suboptimal for the only 3-days of wear of the OmniPod. VetBond is tissue glue, and it takes off the kitty’s hair, and if not careful, takes a layer of skin off the kitty on removal. And a kitty ripping off its sensor on purpose or accident won’t be very careful. Suboptimal.

However, Skinister Medical Adhesive has so far been perfect for what we use it for. I now use Skinister when I want to re-adhere one of my own sensors … seriously awesome stuff in my opinion (this is not sponsored, just my opinion.)

Picture below: A screenshot of Buddy’s Nightscout after a few days on iAPS (previously “FreeAPS-X). Pretty nice euglycemic curve, Mr. Buddy. Your T1D friends and supporters envie you your pretty curve ;-).

The upper curve shows 12 hours of his blood sugar, the lower curve shows 48 hours. Nightscout is awesome. http://www.nightscout.info | #wearenotwaiting

Buddy’s insulin is diluted for safety reasons and for technical reasons.

And for (rather different) safety reasons, the Omnipod and iAPS only “talk/think” U-100 insulin (100 Units per milliliter). This is very important, and I will explain in a later post why we chose a 10% dilution (math), and how we deal with documenting in Nightscout any injections of LONG-ACTING Glargine U-100.

What’s the insulin in the pump?

FIAsp, or Humalog (I have used either). We use it in the pump at 10% strength – one part fast-acting insulin diluted with nine parts diluent for insulin.

Diluted insulin is often used for human babies and toddlers with Type 1 Diabetes – for safety. And sometimes, to make use of an insulin pump feasible. This is a bit complex and errors can have tragic consequences. I’ll write more about how we are dealing with this group of issues.

What about Glargine?

Insulin Glargine (Lantus) does not belong into a pump – it’s long-acting. For safe and effective insulin infusion you need fast-acting insulin. (more about that later)

Most importantly, GLARGINE / Lantus® INSULIN MUST NEVER BE DILUTED – it works as a long-acting insulin because of its pH. If you dilute glargine, it may work like a short acting insulin – and that is very very bad – too much insulin action at once. Please don’t kill your cat.

We use Insulin Glargine sometimes instead of an Omnipod. This is to inject long acting Insulin which lasts the kitty for several hours – instead of an Omnipod filled with short-acting insulin.

The time the Glargine is working depends on the dose, and other things. Buddy was not making enough of his own insulin., we can see in his blood sugar curves how long the insulin was active: When given 2.0 Units of Glargine, we could see its effect on the blood sugar last about 8 hours, after that, his BG went up quickly. In the picture below he had 3 Units of Glargine – which was too much, and the Glargine pushed his blood sugar too low.

I mentioned human error a few times. My biggest concern is that one day I accidentally give Buddy undiluted fast-acting insulin, instead of his diluted insulin. To reduce that risk, I always prepare for Buddy a vial of 10% diluted U-100 insulin (making it U-10 insulin). I clearly mark his bottle of insulin with colorful tapes – visual and tactile markers for me.

Should we go out of town and a petsitter needs to come to the house and inject him, I will move all of my own insulin into a different fridge. Better paranoid than sorry. 1 unit of fast-acting insulin would drop his blood sugar by approximately 1,000 mg/dl, according to his most recent insulin sensitivity factor used in his closed-loop system. Not much margin for error!!!

Following that, Buddy’s IOB in this picture is 0.065 Units, and his temporary basal rate is 0.03 units/hour.

So, now, we’re back to chilling in the Scratchy Box.

Cheers!

Link to one of the useful-for-me papers:

In those first days, I read quite a few veterinary papers and websites. Much, but not everything I read, I found useful. I’ll add links to the home page when I figure out how to do that, but here is one paper about implanting an insulin pump controlled by a smartphone app into an uncooperative diabetic kitty. The cat went into remission within a month. I imagine she’s a Tortoiseshell kitty 😉 … (feisty and uncooperative – sounds like my own Tortie SweetPea)

• When they removed the pump, the surgeon accidentally cut into the pump’s reservoir – and the U400 insulin leaked into the kitty.
• All is well that ends well – Apparently, after a day on i.v. glucose infusion, the kitty was alright. Phhhew.