Never even heard of the 50/50 thing until relatively recently (dx'd in 1983). I started on a pump while under care of Joslin and they just had me do the usual fasting/testing thing (pre-CGM). Over time I've tweaked my settings when fasting levels indicated the need, and of course now I use an AID pump (Tandem CIQ). Thing is, I eat restricted carb, so my basal / bolus ratio leans much more toward basal just b/c I don't bolus all that much. Currently I'm at ~70/30 basal to bolus, with a 5.3 AIC and 95-100% TIR. I don't see how I'm going to benefit from trying to drop my basal component to a lower percentage than bolus under those conditions.
Yeah, the whole "fasting/testing" thing to determine your basal rate has long been disproven, because when you fast, your liver produces glucose to autoregulate. This had been assumed to happen "all the time", but has since been found that it doesn't happen when you eat. Therefore, for those who consume more than a "low-carb diet", the physiological need for basal insulin drops dramatically.
As for restricted diets, that's essentially putting your liver in an "autoregulate" mode all the time, so your basal rate will be far higher than for those who consume more carbs. Your 70/30 ratio is statistically inline with normal physiology for low-carb. And yes, the low A1c and high TIR also correlate to low-carb diets.
It's too bad impressive glucose control isn't the only factor that determines long-term health outcomes. Metabolic health is a much larger factor than glucose levels.
Hi Dan, I'm curious about the tradeoff between basal and sleep quality. As other commenters have pointed out, and as your article shows as well, you take zero basal units but either still experience overnight hypos OR pre-dose carbs to avoid it. Both of these would seem to impact sleep quality - you're waking up to take glucose or you're waking up to dose to come down after over-eating. Even if you successfully pre-carb and don't have to wake up, the digestion process is impacting your sleep as well, right?
My question is: given what we know about how important sleep quality is for long-term health and reduction in all-cause mortality, is this tradeoff worth it? E.g. if you were on a 30% basal ratio and generally experienced no overnight interruptions, would you still try to reduce basal to zero knowing that it would likely increase the amount of overnight interruptions and thus reduce sleep quality overall?
yes, sleep quality is a thing! A big thing! It's admittedly a no-win situation. However, I do feel I've got the lesser of two evils. The frequency and severity of the hypos that I had when I was using ANY basal insulin is far worse than the current state of affairs---much fewer hypos and having to carb bolus now and then. Sleep disruption yes, but it happens either way. Your suggestion of taking 30% basal---or, as I stated above, ANY basal---means that my hypos would return. Again, I have hypos at night, even with no insulin onboard. Zero. Nada. Zilch. I realize the article is long and dense, but it's worth reading again if that part isn't clear.
I agree that sleep disruption is bad, but it's a tradeoff that cannot be avoided.
Now, I will add here--and you ladies may wish to look away for this part--my being on the slippier side of 50 (ok, 60, but who's counting) means that getting up to pee every two hours as matter of physiological reality means that my sleep is already disrupted. The only silver lining to that, and I mean a *very thin* silver lining, is that, since I'm up, I can see what my glucose is doing and intervene as necessary. Believe it or not, this really does help in nocturnal glucose control.
Naturally, I do not recommend that anyone (who can help it) wake up every two hours and watch their CGMs. But if you do, wear blue light glasses, as it helps those dips not look so bad. And if you can't get back to sleep, you could do some light reading, say, of works by this year's Nobel Prize Winner for Literature, Hungarian scribe László Krasznahorkai. In his debut novel, "Satantango", each chapter is a single paragraph with no line breaks. His most recent novel to appear in English, "Herscht 07769", is a single sentence that unfolds over 400 pages.
You think MY articles are long! How cow! Consider yourself lucky.
Where was I?
Oh yeah, sleep disruption. I think you can tell by this response that I either had a *fantastic* night's sleep last night, or a completely horrible one, leaving my brain in a fog. If you can't tell, you're not alone. We'll leave that for another article.
But I will say, all things considered, despite the fact that I DO get up every two hours, I actually get sufficient sleep most nights. My Oura ring *and* my apple watch, which, by the way often disagree about a lot of things, each report that I get 7.5hrs of sleep each night, with just the right amount of deep vs. REM sleep.
I've stated in another comment that nocturnal glucose control is *the* hardest part of my particular T1D management. I do not see an easy solution here. But I do get in a lot of light reading.
Informative article and great insights. I appreciate the links to other resources.
This really highlights that the 50/50 basal rule (and other similar insulin rules for that matter) should be taken with a grain of salt.
I have to be honest, the no-basal approach was a bit surprising. In the July 1st 2020 chart there are correction doses at 2am and 5am, while July 14th 2020 chart has them at 4am and 5am. I often hear that the ideal basal would lead to flat overnight blood sugar without dextrose tabs or correction boluses. Is this approach meant more as an temporary experiment?
And you're prompting me to actually dedicate an entire post to nocturnal hypoglycemia. It's a huge issue, and one that I personally have major struggles with (as evidenced by this last article on basal reduction). I am stumped as to how to avoid it, because I get hypos without ANY insulin onboard. I avoid most of it by pre-carbing before going to bed (nuts, which have fat and protein, which are like a "carb basal"), and that works well. But it's imperfect (cuz nothing is).
And if you're thinking *glucagon drip*, as is being tested in some dual-hormone insulin pumps under development. well, that's also imperfect. Clinical data on glucagon injections alone show wide variability in efficacy among individuals, and within the same person on different occasions. It shows promise, but it's still too early in development. I touch on glucagon injection variability in my article on how to treat hypos: https://danheller.substack.com/p/the-best-way-to-treat-hypoglycemia
But now that you have me thinking more deeply about hypos, I just uploaded a ton of recent CGM+carb+insulin data to ChatGPT and had it produce a report analyzing all this data, and I'm finding some surprising facts that I sort of intuited, but had never fully quantitated and analyzed. I'm still in the process, but here's a teaser:
In the past 90 days, I've had 111 nighttime carb events (12–5 AM) that I had to do in order to treat or avert hypos. 22 of these (≈20%) occurred during insulin-free gaps, which is defined as starting 5 hours AFTER my last bolus (usually around 7-8pm). These events ranged from about 1.7 hrs to 6+ hrs INSIDE THE GAP-- that is, I still had hypos UP TO 6+ hours AFTER the end of the 5 hour period since the last bolus. That would include my having a hypo at 5-6am (or, potentially, having to eat carbs to avert it). More shockingly, there were 2.18 events per night (on nights where at least one event happened). So, it's not like I just get a hypo and treat it. When they happen, they usually happen a lot. My guess is that they happen when I under-estimate the fat+protein before bed. When I dose that well, the night is generally pretty smooth.
That leads to your observation of my bolusing at 4am, and it comes down to this: Carb-dosing compensation for lows (whether to treat or to avert) is an imprecise task. If I overshoot, then my glucose rises and I need a bolus. Making things even more complicated--and something I considered including, but didn't because the article is long enough--the effect of food absorption variability, which I talk about in that same article I linked to before:
Food absorption variability is another stealthy, overlooked phenomenon that T1Ds should be aware of. In the case that you're looking at, chances are very likely that my food absorption was variable to the point where the glucose averted the hypo, but then the absorption paused and restarted. As glucose went up, I had to bolus. I'm looking for more data like that in the analytics that I'm having chatgpt perform.
You can find many more charts of mine in my article on CGMs, where I illustrate these events in far greater detail.
A hypo after 11 hours since last injection...? That doesn't sound right. Something else is going on, either insulin is severely lagging somehow, or your own leftover cells working? Doubt it can be explaned by "NIMGU" or anything like that. There is some truth to that but this effect is never that heavy, at least for me. Not to mention the more stable your exercise is the more you/body adapts and the effect has even less impact or you need to constantly step up your exercise to infinity for it to be felt. At least, that is my experience.
111 events of hypo (or if it would be hyper) over 90 days just says every night is basically destroyed. Doubt you even need AI to analyse anything and tell you that.
To be honest, all of this is really depressing. Basically it would seem there's no way out. You can tone down your basal/drop it completely but you'll need to be way more engaged, it's just a tradeoff. Would be a nice study showing which wears you down more -- "overbasalisation" and less stress by tuning out, or trying to perfectly tune your insulin needs but needing way more attention. Fortunately, it's not as binary as it would seem.
Another thing is that way different dynamics come into play when I have a flat profile of Tresiba than being completely off basal. I feel like body adapts to the constant flow of insulin somehow and has it easier than chaotic drips of insulin based off of CGM readings. Like I would never be able to achieve a night like this without any basal https://www.icloud.com/iclouddrive/08fDLvMLQFIq9aPWOQp3qpdDQ#Screenshot_2025-10-15_at_16.42.45 I even wonder how is this even possible without any engagement (ignore the day, which I could've done better). Have you ever tried Tresiba? Lantus has nasty peaks, I've stopped using it in 2016 or so.
I know at first it doesn't make sense injecting basal when you have hypos without anything but the different dynamics throughout the day might change something. It's just something constant flow does either with Tresiba or a pump constantly microdosing (say Humalog) which also works for me, but because of other issues it brings I like to stay with MDI.
It just seems your way is too hardcore, some people don't even use CGM's, check their BG 3 times a day and have good control, maybe they just say that but still. Who wins at the end of the day, I'd say definitely the one who makes it all much more simple and still succeeds, no basal approach is definitely not simple, at least from my experience after trying it.
I echo that a small dose of Tresiba is worth trying, if possible.
Duration is not the only difference between Tresiba and Lantus, some evidence suggests that Tresiba can result in more predictable absorption. There is even evidence that it can even reduce overnight lows in some type 1 diabetics. More information for Dan and others who are interested in the science:
Thanks for putting together this series. I've been on MDI for 28 years, but am currently exploring trying out hackable Omnipods (with a diy AID supportive Canadian endo and team in the loop for using AAPS). I have all the usual concerns about the pump trade-offs (tissue inflammation from steady insertion over 2-3 days, inability to know for sure how much of your dose really went in - tunnelling, lose catheters, and other mechanical failings) but I don't know that I'm up for waking up to dose short acting in the night, and am intrigued by the diy community driven approaches and playing with that in a semi-closed loop fashion (have zero interest in Omnipod's black box of a hands-off commercial algorithm). I just hope my tissues play nicely with it.
I've been quite happy with Levemir the last 15+ years - I've found its shorter acting time and consistent slight peak to work perfectly with timing for dawn phenomenon, and it has allowed me to reliably sleep peacefully through the night with only the notably rare nocturnal hypo for years. Because it doesn't last as long, I've also found it easy to adapt to days I know I'm exercising more.
Unfortunately, no biosimilars have come onto the market, and Novo Nordisk is abandoning its production in favour of ramping up profitable Wegovy and Ozempic so I've been trying to adapt to Basaglar as the alternative (glargine biosimilar to Lantus) and it's been two weeks of chronic nighttime hypoglycemia with a split dose. I fear glargine may simply not play well with my body and needs. (But the alternative of Tresiba seems even less ideal as it purports to be even flatter in profile and lasts up to 48 hours! Talk about inflexible.)
Fasting is relatively flat during the day, but if I go down a unit at night I'm going high from dawn phenomenon before I can get up. Up just half a unit (I went out and got syringes to draw from the pens and play with half dosing) and I'm crashing low two or three times a night. Tried to take the evening dose earlier with dinner to see if it improved things, but it seemed to make marginal difference. After two sleepless weeks I'm about to try cutting units off the day so that the night dose isn't stacking onto a higher totally daily dose and see if that's the winner. At least until I can try out the pump. I feel a bit unnerved having to take what feels like so many units of lispro without the cushion of basal, but your article has inspired me to be relatively aggressive with cutting the daytime portion of glargine (at least I'm awake to watch the CGM).
Love this. I had to tear myself away for my workout. 💪🏻. (30" weights/30 in the pool) I have to read everything again and take notes because I am living the opposite. I only bolus for things like fresh juices and NA beers. I came to this formula using my own version of active learning and my CGM and started splitting doses (against my endo's advice). I rarely go low during exercise up to 90 minutes and have a pretty good method for stretching things out up to 3+ hours. I have always blamed my exercise lows on the Basaglar. If I don't have any OB, I do not experience sharp drops. I am VERY curious to see what happens if I flip things. I will say that I misunderstood your article on DKA - my impression was that keeping basal in the body prevents DKA, but it sounds like you are saying something different here. I need to read it again.
My next article gets into the weeds on exercise: A primer for T1Ds, and it'll cover a lot of your questions. I'm almost done, so stay tuned.
Your basal/bolus description has no detail at all, so it's hard to comment on it, but I'll add this: your metabolic profile is in great shape given your exercise regimen--both cardio and anaerobic training. That's what gets you the level of glucose and insulin efficiency that often eludes others, and is probably what allows you to achieve the balance you've crafted for yourself. Your muscles are a glucose sink, and your insulin sensitivity allows a great deal of flexibility that others might not get.
Unexpected hypos are often due to excess insulin, of course, and basal is usually the wildcard because you never know how much you *really* have onboard, making all other dosing decisions subject to a degree of error. Reducing basal reduces the error, but also requires you to be more diligent about bolusing the right amounts in a timely manner. I suppose that's the case even if you do have a high basal ratio, so it's work one way or another. You pick your own adventure.
Regarding DKA and basal -- your recollection needs a small tweak: The *original* assumption was that keeping [high] basal onboard prevents DKA, but that has been debunked, and (as you've read in my last three articles), insulin-free gaps are physiologically normal, especially for those who engage in exercise.
I actually "got into" a bit of an online tif over a post that talked about going 5 hours during the NY Marathon without bolusing for any of the carbs she ate. One of the commenters LOST HER MIND over it. Screaming DKA and such. There is so much ignorance from people who do not exercise to that level. I was dx'd in 1982 at 13. We were clueless as our numbers. we just ate and dosed and hoped for a good a1c, so I just continued my athletic life as before.
I use a total of about 11 to 13 g of insulin in bolus doses daily. This varies of course, but that’s generally where I am. I try not to take more than three units per meal, to help keep me a little more even.
My blood sugars seem to vary significantly day-to-day. Some days I feel like I’m going higher than I would like, maybe 160 or so, a couple hours after high-protein meal. Other days, like today, I seem to be running on the low side all day! And my diet is very similar day to day. Which is why I think the variation is related to exercise. Sometimes that effect seems to kick in within an hour, other times it seems to take over 24 hours.
I’m the type of person who is constantly looking at their GCM anyhow, so this comes quite naturally to me. I would prefer to actually forget about it more often!
Let me know if you have any suggestions. Thanks so much!
Thanks for this article! I was diagnosed with Type 1 eight years ago, at age 49. I have struggled with exercise, previously doing a lot of running. I am more focused on strength training now, and high intensity interval training (which is better for me anyhow, though i really miss long distance running!). I am currently taking 3 units of Tresiba, which I recently decreased from 4, in an attempt to fix frequent hypos (often down to 40). After reading this, I am tempted to decrease it even more, as I often have to avoid even going for a walk in the late afternoon, or after dinner, as I know I will drop like a rock. However, I am a bit worried about going too high... Have you found that you are taking more insulin with meals to avoid high blood sugars? I am also interested in your thoughts on dosing for protein...
Jessica -- you're very similar to me, insofar as the exercise routine goes. You didn't say what ratio your 3u of tresiba is relative to your TDD, but given your description of frequent lows, especially in the 40s, it makes sense to be more aggressive. As you suggested, paring back on tresiba further will reveal a lot more about your true metabolic profile.
I feel that any individual that is motivated and capable enough to pay super-close attention to their CGM is also capable of paring back on basal and truly fine-tuning their management without taking on any health risks. Keep us/me posted and IM me if you want to take this offline.
Thanks for this in-depth article. You mention hypos many times but I don't see anywhere what blood sugar level is defined as a hypo in your experience. Thanks for helping me understand this.
You know, it didn't even occur to me that there aren't graphs showing BG levels dropping into the 40s and 50s.. I made all those charts back in 2020 when I first had this whole experience, and you can be sure there were many such nights. The data is long gone, so I can't reproduce new charts now.
For context, I made these charts as part of a larger document that I showed to my endo to demonstrate the larger picture of what was happening to me and to help figure out what was going on. Those particular graphs in the article are from that document, and what they illustrate is the repetition of my BG falling rapidly, recovering when I ate dextrose, and then rapidly falling again.
You can easily infer that, had I not intervened, that those drops would have continued into dangerous territory. The first graphic illustrated a night with three large interventions, all without any (fast-acting) insulin. I had Lantus onboard, which I didn't realize at the time was having this effect.
the other graph illustrates a similar pattern, but less extreme.
Thanks for pointing that out! I now wish I had that data and I can show the nightmare I was living through. But my guess is that it's nothing all of us haven't experienced.
Ahhhh thanks for that. I totally misinterpreted the graphs you showed thinking that you took the glucose for a "stable" level of around 80 and not a level that was continuing on down. Yikes! Thanks for setting me straight. That must have been awful.
Your best article yet, IMHO, as you pull together the previous articles. My observations as a primary care doc…. many many T1D’s are still on too much insulin. Your program will work quite well for a T1D who is willing to get very involved in their management. Even for those not so inclined, there would seem to be benefit even in reducing basal dosing to 25% of TDI. Keep up your great work!
Rapid Acting Insulin Analogs (Insulin Aspart, Insulin Lispro, Insulin Glulisine) which have an onset of action of 5 to 15 minutes, peak effect in 1 to 2 hours and duration of action that lasts 4-6 hours. With all doses, large and small, the onset of action and the time to peak effect is similar. The duration of insulin action is, however, affected by the dose – so a few units may last 4 hours or less, while 25 or 30 units may last 5 to 6 hours. As a general rule, assume that these insulins have a duration of action of 4 hours.
*************************************
>>>>>>>>>>Note that even the rapid acting insulins will have a "bolus-basal effect", increasingly so with larger doses; and multiply times meals per day.
II) For those of us who continue to use a Long-Acting basal insulin: to regularly evaluate one's dosage of basal insulin, it works very well to fast from after dinner the night before until dinner the following day (20 or so hours), while watching blood glucose levels closely (Dr. Richard K. Bernstein gives explicit instructions in his book.). The short fast may have other benefits, yes, Dr. Jason Fung's book is on my reading list.
And a bonus 3rd point: Going very-low-carb reduces one's need for basal and bolus insulins.
Thanks again for this post and bringing some new things to my attention.
Just trying to help- as your blog reminds us, It's Just Not That Simple.
Janice -- My article on "Insulin absorption variability" talks about the topic in more depth than simple absorption curves, peaks and taper. Lipodystrophy is a big deal that many are unaware of.
Yes, dose size matters a lot--which is why I never take more than 5u in a single injection. See the article for details.
The problem with using rapid acting insulins at night is the immediacy of effect, rather than the duration, as you pointed out. You can't get both short duration and blunted peaks at the same time in a single insulin formula. That said, a good bolus of aspart in the morning takes care of the dawn effect quite nicely for me.
Never even heard of the 50/50 thing until relatively recently (dx'd in 1983). I started on a pump while under care of Joslin and they just had me do the usual fasting/testing thing (pre-CGM). Over time I've tweaked my settings when fasting levels indicated the need, and of course now I use an AID pump (Tandem CIQ). Thing is, I eat restricted carb, so my basal / bolus ratio leans much more toward basal just b/c I don't bolus all that much. Currently I'm at ~70/30 basal to bolus, with a 5.3 AIC and 95-100% TIR. I don't see how I'm going to benefit from trying to drop my basal component to a lower percentage than bolus under those conditions.
Yeah, the whole "fasting/testing" thing to determine your basal rate has long been disproven, because when you fast, your liver produces glucose to autoregulate. This had been assumed to happen "all the time", but has since been found that it doesn't happen when you eat. Therefore, for those who consume more than a "low-carb diet", the physiological need for basal insulin drops dramatically.
This is covered in my article on physiology and basal needs: https://danheller.substack.com/p/basal-dosing-how-much-do-we-need
As for restricted diets, that's essentially putting your liver in an "autoregulate" mode all the time, so your basal rate will be far higher than for those who consume more carbs. Your 70/30 ratio is statistically inline with normal physiology for low-carb. And yes, the low A1c and high TIR also correlate to low-carb diets.
It's too bad impressive glucose control isn't the only factor that determines long-term health outcomes. Metabolic health is a much larger factor than glucose levels.
https://danheller.substack.com/p/the-paradox-of-low-carb-diets-a1c-vs-metabolic-health
Hi Dan, I'm curious about the tradeoff between basal and sleep quality. As other commenters have pointed out, and as your article shows as well, you take zero basal units but either still experience overnight hypos OR pre-dose carbs to avoid it. Both of these would seem to impact sleep quality - you're waking up to take glucose or you're waking up to dose to come down after over-eating. Even if you successfully pre-carb and don't have to wake up, the digestion process is impacting your sleep as well, right?
My question is: given what we know about how important sleep quality is for long-term health and reduction in all-cause mortality, is this tradeoff worth it? E.g. if you were on a 30% basal ratio and generally experienced no overnight interruptions, would you still try to reduce basal to zero knowing that it would likely increase the amount of overnight interruptions and thus reduce sleep quality overall?
yes, sleep quality is a thing! A big thing! It's admittedly a no-win situation. However, I do feel I've got the lesser of two evils. The frequency and severity of the hypos that I had when I was using ANY basal insulin is far worse than the current state of affairs---much fewer hypos and having to carb bolus now and then. Sleep disruption yes, but it happens either way. Your suggestion of taking 30% basal---or, as I stated above, ANY basal---means that my hypos would return. Again, I have hypos at night, even with no insulin onboard. Zero. Nada. Zilch. I realize the article is long and dense, but it's worth reading again if that part isn't clear.
I agree that sleep disruption is bad, but it's a tradeoff that cannot be avoided.
Now, I will add here--and you ladies may wish to look away for this part--my being on the slippier side of 50 (ok, 60, but who's counting) means that getting up to pee every two hours as matter of physiological reality means that my sleep is already disrupted. The only silver lining to that, and I mean a *very thin* silver lining, is that, since I'm up, I can see what my glucose is doing and intervene as necessary. Believe it or not, this really does help in nocturnal glucose control.
Naturally, I do not recommend that anyone (who can help it) wake up every two hours and watch their CGMs. But if you do, wear blue light glasses, as it helps those dips not look so bad. And if you can't get back to sleep, you could do some light reading, say, of works by this year's Nobel Prize Winner for Literature, Hungarian scribe László Krasznahorkai. In his debut novel, "Satantango", each chapter is a single paragraph with no line breaks. His most recent novel to appear in English, "Herscht 07769", is a single sentence that unfolds over 400 pages.
You think MY articles are long! How cow! Consider yourself lucky.
Where was I?
Oh yeah, sleep disruption. I think you can tell by this response that I either had a *fantastic* night's sleep last night, or a completely horrible one, leaving my brain in a fog. If you can't tell, you're not alone. We'll leave that for another article.
But I will say, all things considered, despite the fact that I DO get up every two hours, I actually get sufficient sleep most nights. My Oura ring *and* my apple watch, which, by the way often disagree about a lot of things, each report that I get 7.5hrs of sleep each night, with just the right amount of deep vs. REM sleep.
I've stated in another comment that nocturnal glucose control is *the* hardest part of my particular T1D management. I do not see an easy solution here. But I do get in a lot of light reading.
Informative article and great insights. I appreciate the links to other resources.
This really highlights that the 50/50 basal rule (and other similar insulin rules for that matter) should be taken with a grain of salt.
I have to be honest, the no-basal approach was a bit surprising. In the July 1st 2020 chart there are correction doses at 2am and 5am, while July 14th 2020 chart has them at 4am and 5am. I often hear that the ideal basal would lead to flat overnight blood sugar without dextrose tabs or correction boluses. Is this approach meant more as an temporary experiment?
Excellent observation!
And you're prompting me to actually dedicate an entire post to nocturnal hypoglycemia. It's a huge issue, and one that I personally have major struggles with (as evidenced by this last article on basal reduction). I am stumped as to how to avoid it, because I get hypos without ANY insulin onboard. I avoid most of it by pre-carbing before going to bed (nuts, which have fat and protein, which are like a "carb basal"), and that works well. But it's imperfect (cuz nothing is).
And if you're thinking *glucagon drip*, as is being tested in some dual-hormone insulin pumps under development. well, that's also imperfect. Clinical data on glucagon injections alone show wide variability in efficacy among individuals, and within the same person on different occasions. It shows promise, but it's still too early in development. I touch on glucagon injection variability in my article on how to treat hypos: https://danheller.substack.com/p/the-best-way-to-treat-hypoglycemia
But now that you have me thinking more deeply about hypos, I just uploaded a ton of recent CGM+carb+insulin data to ChatGPT and had it produce a report analyzing all this data, and I'm finding some surprising facts that I sort of intuited, but had never fully quantitated and analyzed. I'm still in the process, but here's a teaser:
In the past 90 days, I've had 111 nighttime carb events (12–5 AM) that I had to do in order to treat or avert hypos. 22 of these (≈20%) occurred during insulin-free gaps, which is defined as starting 5 hours AFTER my last bolus (usually around 7-8pm). These events ranged from about 1.7 hrs to 6+ hrs INSIDE THE GAP-- that is, I still had hypos UP TO 6+ hours AFTER the end of the 5 hour period since the last bolus. That would include my having a hypo at 5-6am (or, potentially, having to eat carbs to avert it). More shockingly, there were 2.18 events per night (on nights where at least one event happened). So, it's not like I just get a hypo and treat it. When they happen, they usually happen a lot. My guess is that they happen when I under-estimate the fat+protein before bed. When I dose that well, the night is generally pretty smooth.
That leads to your observation of my bolusing at 4am, and it comes down to this: Carb-dosing compensation for lows (whether to treat or to avert) is an imprecise task. If I overshoot, then my glucose rises and I need a bolus. Making things even more complicated--and something I considered including, but didn't because the article is long enough--the effect of food absorption variability, which I talk about in that same article I linked to before:
https://danheller.substack.com/p/the-best-way-to-treat-hypoglycemia
Food absorption variability is another stealthy, overlooked phenomenon that T1Ds should be aware of. In the case that you're looking at, chances are very likely that my food absorption was variable to the point where the glucose averted the hypo, but then the absorption paused and restarted. As glucose went up, I had to bolus. I'm looking for more data like that in the analytics that I'm having chatgpt perform.
You can find many more charts of mine in my article on CGMs, where I illustrate these events in far greater detail.
https://danheller.substack.com/p/the-dexcom-g7-vs-g6-which-is-better
A hypo after 11 hours since last injection...? That doesn't sound right. Something else is going on, either insulin is severely lagging somehow, or your own leftover cells working? Doubt it can be explaned by "NIMGU" or anything like that. There is some truth to that but this effect is never that heavy, at least for me. Not to mention the more stable your exercise is the more you/body adapts and the effect has even less impact or you need to constantly step up your exercise to infinity for it to be felt. At least, that is my experience.
111 events of hypo (or if it would be hyper) over 90 days just says every night is basically destroyed. Doubt you even need AI to analyse anything and tell you that.
To be honest, all of this is really depressing. Basically it would seem there's no way out. You can tone down your basal/drop it completely but you'll need to be way more engaged, it's just a tradeoff. Would be a nice study showing which wears you down more -- "overbasalisation" and less stress by tuning out, or trying to perfectly tune your insulin needs but needing way more attention. Fortunately, it's not as binary as it would seem.
Another thing is that way different dynamics come into play when I have a flat profile of Tresiba than being completely off basal. I feel like body adapts to the constant flow of insulin somehow and has it easier than chaotic drips of insulin based off of CGM readings. Like I would never be able to achieve a night like this without any basal https://www.icloud.com/iclouddrive/08fDLvMLQFIq9aPWOQp3qpdDQ#Screenshot_2025-10-15_at_16.42.45 I even wonder how is this even possible without any engagement (ignore the day, which I could've done better). Have you ever tried Tresiba? Lantus has nasty peaks, I've stopped using it in 2016 or so.
I know at first it doesn't make sense injecting basal when you have hypos without anything but the different dynamics throughout the day might change something. It's just something constant flow does either with Tresiba or a pump constantly microdosing (say Humalog) which also works for me, but because of other issues it brings I like to stay with MDI.
It just seems your way is too hardcore, some people don't even use CGM's, check their BG 3 times a day and have good control, maybe they just say that but still. Who wins at the end of the day, I'd say definitely the one who makes it all much more simple and still succeeds, no basal approach is definitely not simple, at least from my experience after trying it.
I echo that a small dose of Tresiba is worth trying, if possible.
Duration is not the only difference between Tresiba and Lantus, some evidence suggests that Tresiba can result in more predictable absorption. There is even evidence that it can even reduce overnight lows in some type 1 diabetics. More information for Dan and others who are interested in the science:
https://jamanetwork.com/journals/jama/fullarticle/2635629
https://pubmed.ncbi.nlm.nih.gov/22594461/
Thanks for putting together this series. I've been on MDI for 28 years, but am currently exploring trying out hackable Omnipods (with a diy AID supportive Canadian endo and team in the loop for using AAPS). I have all the usual concerns about the pump trade-offs (tissue inflammation from steady insertion over 2-3 days, inability to know for sure how much of your dose really went in - tunnelling, lose catheters, and other mechanical failings) but I don't know that I'm up for waking up to dose short acting in the night, and am intrigued by the diy community driven approaches and playing with that in a semi-closed loop fashion (have zero interest in Omnipod's black box of a hands-off commercial algorithm). I just hope my tissues play nicely with it.
I've been quite happy with Levemir the last 15+ years - I've found its shorter acting time and consistent slight peak to work perfectly with timing for dawn phenomenon, and it has allowed me to reliably sleep peacefully through the night with only the notably rare nocturnal hypo for years. Because it doesn't last as long, I've also found it easy to adapt to days I know I'm exercising more.
Unfortunately, no biosimilars have come onto the market, and Novo Nordisk is abandoning its production in favour of ramping up profitable Wegovy and Ozempic so I've been trying to adapt to Basaglar as the alternative (glargine biosimilar to Lantus) and it's been two weeks of chronic nighttime hypoglycemia with a split dose. I fear glargine may simply not play well with my body and needs. (But the alternative of Tresiba seems even less ideal as it purports to be even flatter in profile and lasts up to 48 hours! Talk about inflexible.)
Fasting is relatively flat during the day, but if I go down a unit at night I'm going high from dawn phenomenon before I can get up. Up just half a unit (I went out and got syringes to draw from the pens and play with half dosing) and I'm crashing low two or three times a night. Tried to take the evening dose earlier with dinner to see if it improved things, but it seemed to make marginal difference. After two sleepless weeks I'm about to try cutting units off the day so that the night dose isn't stacking onto a higher totally daily dose and see if that's the winner. At least until I can try out the pump. I feel a bit unnerved having to take what feels like so many units of lispro without the cushion of basal, but your article has inspired me to be relatively aggressive with cutting the daytime portion of glargine (at least I'm awake to watch the CGM).
Love this. I had to tear myself away for my workout. 💪🏻. (30" weights/30 in the pool) I have to read everything again and take notes because I am living the opposite. I only bolus for things like fresh juices and NA beers. I came to this formula using my own version of active learning and my CGM and started splitting doses (against my endo's advice). I rarely go low during exercise up to 90 minutes and have a pretty good method for stretching things out up to 3+ hours. I have always blamed my exercise lows on the Basaglar. If I don't have any OB, I do not experience sharp drops. I am VERY curious to see what happens if I flip things. I will say that I misunderstood your article on DKA - my impression was that keeping basal in the body prevents DKA, but it sounds like you are saying something different here. I need to read it again.
My next article gets into the weeds on exercise: A primer for T1Ds, and it'll cover a lot of your questions. I'm almost done, so stay tuned.
Your basal/bolus description has no detail at all, so it's hard to comment on it, but I'll add this: your metabolic profile is in great shape given your exercise regimen--both cardio and anaerobic training. That's what gets you the level of glucose and insulin efficiency that often eludes others, and is probably what allows you to achieve the balance you've crafted for yourself. Your muscles are a glucose sink, and your insulin sensitivity allows a great deal of flexibility that others might not get.
Unexpected hypos are often due to excess insulin, of course, and basal is usually the wildcard because you never know how much you *really* have onboard, making all other dosing decisions subject to a degree of error. Reducing basal reduces the error, but also requires you to be more diligent about bolusing the right amounts in a timely manner. I suppose that's the case even if you do have a high basal ratio, so it's work one way or another. You pick your own adventure.
Regarding DKA and basal -- your recollection needs a small tweak: The *original* assumption was that keeping [high] basal onboard prevents DKA, but that has been debunked, and (as you've read in my last three articles), insulin-free gaps are physiologically normal, especially for those who engage in exercise.
I actually "got into" a bit of an online tif over a post that talked about going 5 hours during the NY Marathon without bolusing for any of the carbs she ate. One of the commenters LOST HER MIND over it. Screaming DKA and such. There is so much ignorance from people who do not exercise to that level. I was dx'd in 1982 at 13. We were clueless as our numbers. we just ate and dosed and hoped for a good a1c, so I just continued my athletic life as before.
Hi Dan,
I use a total of about 11 to 13 g of insulin in bolus doses daily. This varies of course, but that’s generally where I am. I try not to take more than three units per meal, to help keep me a little more even.
My blood sugars seem to vary significantly day-to-day. Some days I feel like I’m going higher than I would like, maybe 160 or so, a couple hours after high-protein meal. Other days, like today, I seem to be running on the low side all day! And my diet is very similar day to day. Which is why I think the variation is related to exercise. Sometimes that effect seems to kick in within an hour, other times it seems to take over 24 hours.
I’m the type of person who is constantly looking at their GCM anyhow, so this comes quite naturally to me. I would prefer to actually forget about it more often!
Let me know if you have any suggestions. Thanks so much!
Thanks for this article! I was diagnosed with Type 1 eight years ago, at age 49. I have struggled with exercise, previously doing a lot of running. I am more focused on strength training now, and high intensity interval training (which is better for me anyhow, though i really miss long distance running!). I am currently taking 3 units of Tresiba, which I recently decreased from 4, in an attempt to fix frequent hypos (often down to 40). After reading this, I am tempted to decrease it even more, as I often have to avoid even going for a walk in the late afternoon, or after dinner, as I know I will drop like a rock. However, I am a bit worried about going too high... Have you found that you are taking more insulin with meals to avoid high blood sugars? I am also interested in your thoughts on dosing for protein...
Jessica -- you're very similar to me, insofar as the exercise routine goes. You didn't say what ratio your 3u of tresiba is relative to your TDD, but given your description of frequent lows, especially in the 40s, it makes sense to be more aggressive. As you suggested, paring back on tresiba further will reveal a lot more about your true metabolic profile.
I feel that any individual that is motivated and capable enough to pay super-close attention to their CGM is also capable of paring back on basal and truly fine-tuning their management without taking on any health risks. Keep us/me posted and IM me if you want to take this offline.
Thanks for this in-depth article. You mention hypos many times but I don't see anywhere what blood sugar level is defined as a hypo in your experience. Thanks for helping me understand this.
You know, it didn't even occur to me that there aren't graphs showing BG levels dropping into the 40s and 50s.. I made all those charts back in 2020 when I first had this whole experience, and you can be sure there were many such nights. The data is long gone, so I can't reproduce new charts now.
For context, I made these charts as part of a larger document that I showed to my endo to demonstrate the larger picture of what was happening to me and to help figure out what was going on. Those particular graphs in the article are from that document, and what they illustrate is the repetition of my BG falling rapidly, recovering when I ate dextrose, and then rapidly falling again.
You can easily infer that, had I not intervened, that those drops would have continued into dangerous territory. The first graphic illustrated a night with three large interventions, all without any (fast-acting) insulin. I had Lantus onboard, which I didn't realize at the time was having this effect.
the other graph illustrates a similar pattern, but less extreme.
Thanks for pointing that out! I now wish I had that data and I can show the nightmare I was living through. But my guess is that it's nothing all of us haven't experienced.
Ahhhh thanks for that. I totally misinterpreted the graphs you showed thinking that you took the glucose for a "stable" level of around 80 and not a level that was continuing on down. Yikes! Thanks for setting me straight. That must have been awful.
Dan,
Your best article yet, IMHO, as you pull together the previous articles. My observations as a primary care doc…. many many T1D’s are still on too much insulin. Your program will work quite well for a T1D who is willing to get very involved in their management. Even for those not so inclined, there would seem to be benefit even in reducing basal dosing to 25% of TDI. Keep up your great work!
Best,
David LaHue
thanks, david -- It's always good hearing from physicians on these topics, as it helps reinforce the concepts to the T1D community.
October 7, 2025
Thanks for this series on overbasalization. Your drive to find answers has worked for you, and good for you for sharing with the rest of us.
There are two points which I believe benefit from more emphasis:
I) Excellent summary from https://diabetesteachingcenter.ucsf.edu/about-diabetes/type-2-diabetes/types-insulin-use-type-2-diabetes :
*********************************
Rapid Acting Insulin Analogs (Insulin Aspart, Insulin Lispro, Insulin Glulisine) which have an onset of action of 5 to 15 minutes, peak effect in 1 to 2 hours and duration of action that lasts 4-6 hours. With all doses, large and small, the onset of action and the time to peak effect is similar. The duration of insulin action is, however, affected by the dose – so a few units may last 4 hours or less, while 25 or 30 units may last 5 to 6 hours. As a general rule, assume that these insulins have a duration of action of 4 hours.
*************************************
>>>>>>>>>>Note that even the rapid acting insulins will have a "bolus-basal effect", increasingly so with larger doses; and multiply times meals per day.
II) For those of us who continue to use a Long-Acting basal insulin: to regularly evaluate one's dosage of basal insulin, it works very well to fast from after dinner the night before until dinner the following day (20 or so hours), while watching blood glucose levels closely (Dr. Richard K. Bernstein gives explicit instructions in his book.). The short fast may have other benefits, yes, Dr. Jason Fung's book is on my reading list.
And a bonus 3rd point: Going very-low-carb reduces one's need for basal and bolus insulins.
Thanks again for this post and bringing some new things to my attention.
Just trying to help- as your blog reminds us, It's Just Not That Simple.
Janice -- My article on "Insulin absorption variability" talks about the topic in more depth than simple absorption curves, peaks and taper. Lipodystrophy is a big deal that many are unaware of.
https://danheller.substack.com/p/the-insulin-absorption-roller-coaster
Yes, dose size matters a lot--which is why I never take more than 5u in a single injection. See the article for details.
The problem with using rapid acting insulins at night is the immediacy of effect, rather than the duration, as you pointed out. You can't get both short duration and blunted peaks at the same time in a single insulin formula. That said, a good bolus of aspart in the morning takes care of the dawn effect quite nicely for me.