Tresiba vs. Lantus study: much less night-time hypoglycemia

Major study shows that, Tresiba (insulin degludec), compared to Lantus (insulin glargine), generates less overall episodes of hypoglycemia, and a bit more than half as many night-time episodes of hypoglycemia (the most dangerous). Significantly fewer patients on Tresiba experience severe hypoglycemia compared to Lantus. Trials were run for both Type 1 and Type 2 diabetes patients.

It occurs to me we never quoted the results of the SWITCH 1 & 2 trials on the forum! There were published on July 4, 2017 on JAMA. So here goes.

Novo Nordisk ran these trials between 2014 and 2016, in about 90 different locations, with a total of about 1,200 patients. The two trials were double-blind, randomized crossover trials:

  • double-blind means that neither the patients nor the medical staff know who is taking what drug;
  • randomized means that the patients are randomly assigned a trial group;
  • crossover means that there are two phases, and that each batch of patients runs through both drugs, one in each phase.

These types of trials establish causality, and are pretty good at ruling out many sources of bias and error: they are the best type of trials you can run!

SWITCH 1 focused on patients with Type 1 diabetes. SWITCH 2 focused on patients with Type 2 diabetes. The results of both trials were similar. In both cases:

  • the overall rate of hypoglycemia was lower for Tresiba (2201/100PYE [Person year’s exposure] vs 2463/100PYE for Lantus in SWITCH 1, 186 vs 265 in SWITCH 2);

  • the night-time rate of hypoglycemia was significantly lower for Tresiba (277/100PYE vs 468/100PYE for Lantus in SWITCH 1, 55 vs 94 in SWITCH 20)

  • fewer patients on Tresiba experienced severe hypoglycemia (10% vs 17% on Lantus in SWITCH 1, 1.6% vs 2.4% in SWITCH 2).

The SWITCH 1 trial write-up in JAMA:

The SWITCH 2 trial write-up in JAMA:


Very interesting results. I hadn’t heard of these trials, so I appreciate you bringing them forward. Additionally, it makes sense that Type 2’s would have less hypo events, but it is shocking to me that it would be a 10 fold difference.


@Michel Thank you for pointing out these important studies.

The results are much of what I would have expected, given the dynamics of the two drugs.

The degludac and the glargine were given as once daily injections, randomized morning or evening. But from my experience (and many others) glargine does not have a true 24 hour activity in many cases. It also does not exhibit a flat activity profile, having peaks and valleys during its activity. Once a day titration can easily lead to overdosage, which in turn would lead to hypoglycemic episodes.

In real life, this is complicated by the fact that a good number of glargine users split the daily dosage in half, taking injections both morning and evening.

Degludac, on the other hand, truly is a 24 hour insulin with no peaks or valleys. Because it is easier to titrate because of this, there is less likelihood of overdosage, hence less likelihood of hypoglycemic episodes.


Solid logic. Seems a contradiction to me that tresiba works better than Lantus because of no peaks or valleys but some think pumping is best of all because of the ability to introduce infinite peaks and valleys…


Sam, I am really glad that you have found an excellent tool to manage your diabetes. With that said, I do find it interesting that you appear to dogmatically question those that have found success using a toolset different than yours. I think that the survey I initiated shows that we have a tremendous pool of very successful diabetics using a variety of interventional tools and diet approaches. That, shows me that the most important tool we can offer is dedication to keeping a healthy A1c in the face of difficult challenges.

Not intending to dogmatically question anyone… just pointing out that this might be a good option to try even though it’s not immediately apparent to all people…

But to some extent we’re all guilty of that. We tend to be critical of R and NPH for example and consider it dated and archaic protocol even though some people have great success with it…

So finally I must ask: Do those of you who successfully use Tresiba not have variable basal requirements throughout the day? Or do you accommodate those with food and bolus adjustments, or some other way?

Tresiba with Afrezza looks more and more attractive, but how would I manage with a 24-hour flatlining basal if I currently pump with seven different basal rates in a 24-hour period? A split dose of Lantus sure didn’t work any miracles.


For many people tresiba disproves their understanding of requiring a variable basal profile. It did for me.

1 Like

As do we (or did until a few weeks ago when we started on a pump). I cannot say, however, that splitting made any improvement to our night-time lows.


@Beacher I actually do find my basal requirements to be fairly consistent throughout the day, and I am feeding the basal far far less with Tresiba than I have ever done with other basals.

I think that a very big reason for varying basal requirements throughout the day is a function of the imperfection of bolus insulins available. For instance, I have about a five or six hour activity duration for novolog. That would be very easy to misinterpret as a varying basal need if I thought it only lasted for three hours.

Of course, as with all things D, YDMV.

1 Like

Is there an explanation for that? Am I to infer that my need for variable basals is all in my head? Or does Tresiba work differently than injected Lantus or pumped Humalog?

If I set one basal for the entire day with my pump, my numbers would be a complete mess. I know this with such certainty that I won’t even test the hypothesis. (Or maybe I will. Will you visit me in hospital?)


If you are within a 10-hour drive we will! Anything for a fellow experimenter :slight_smile:

I do not quite understand how Tresiba appears to decrease the need for variability in basal rate – but many people report what @docslotnick and @Sam mention.

To add to the anecdotal evidence, at the endo clinic last week, our nurse (she is a T1) mentioned to me that she had been experiencing a lot of night-time lows on her pump, and that she had switched to Tresiba 2 months before, and that she was extremely happy with her results, to the extent that she would certainly not consider switching back. She is in her late 20s and very active but not intensely engaged into sports.

1 Like

The only explanation I can come up with is that some people require highly varied basal profiles with some insulins and don’t with others. That doesn’t mean it’s in your head or that your experiences are in any way wrong-- it just means that something else might be outside of your experiences. The experience of requiring a lot of dosing adjustability with some insulins to make them work well doesn’t necessarily transfer to other insulins with different absorption mechanisms I’d suggest.

1 Like

I can tell you from experience unequivocally that Lantus and Tresiba work quite differently.


I am actually getting ready to greatly simplify my son’s basal profile on his pump from the one the clinic recommended (I tuned it heavily the first few weeks). He has about 6 segments right now, but I thinking of going down to maybe 3, just to experiment.

Steven Ponder has a preference for a flat basal over 24 hours himself.

3 - 4 is easily possible. The hardest ones to change, are when you lowered it for a time to deal with consistent lows.

It would be really interesting to see what a pump like the 670G would do if you don’t eat for a day. That might provide some information about varying basal needs.


I doubt it… I think it’s pretty wel unanimous that to use rapid acting analogs for basal they’ve gotta be adjusted to some extent… their absorption just seems to be influenced by a great many factors. We can see with the APS users that their algorithms adjust their rates a lot from what I’ve seen. I’d never dispute that rapid analogs delivered by a pump require a lot of variation in how they’re pumped to achieve consistent control


Sounds admirably simpler than my own regimen, but it sounds like Mr. Ponder does not experience the dawn phenomenon and the afternoon slump. These are circadian hormonal shifts, and I’m still trying to get my head around how Tresiba deals with them.

1 Like

Would probably take a few samples to actually get ones head wrapped around it. It’s likely not something that can be explained in the context of variable basal needs… for some people it just replaces that concept.