What’s the best CGM solution? A simple question with a complex answer…

What’s the best CGM solution? An easy question with a more difficult answer…
What’s the best CGM solution? An easy question with a more difficult answer…

Over the past couple of months, I’ve seen this question asked multiple times on many fora. Facebook, email lists and forums have all seen people looking at CGM asking the same question, or a variation of the question. Which manufacturer’s solution is best? And the answer from the crowd seems to be invariably “Go look at Dexcom”.

But is Dexcom really the best CGM solution FOR YOU? That’s a very different question.

As a precursor, I’ve used the Abbott Freestyle Libre, Dexcom G4, Dexcom G5, Medtronic Enlites with Guardian 2 (release 1 & 2) and with the Minilink transmitter. I’ve not tried the Guardian Connect solution, and as a result, I’m not sure the answer is as straightforward as “Which is best?”.

So let’s take a step back and look at the usual collective answer, which seems, invariably, to be Dexcom. Reversing out of this, one might ask the question, which Dexcom and why?

Accuracy and Longevity

When people answer the question, the two things that are normally foremost in their mind are “How accurate is it?” and “How much does it cost?”. But we should also put the context of the answers in place. Individual experience plays a huge part in the response. This is why the system matters.

If you ask in many of the groups on Facebook, where it often comes down to Medtronic v Dexcom, the experience of both is, for the most part, Medtronic Enlite Sensors with Minilink transmitter (US users have not been able to use the Guardian 2 with their pumps as this, up until recently, has only been available with the European/WW pumps) versus a mixture of G4 and G5 systems from Dexcom. In both cases, the sensors don’t change between Dexcom G4 and G5 and Medtronic current Minilink users and Guardian users. The Enlite sensors were released in 2014 and updated in early 2015, after many people used a Medtronic system, ditched it, and moved on to Dexcom.

Why is this data important? Because it plays a key part in understanding some of the biases that may be prevalent in the discussions taking place.

Various people have done tests looking at the accuracy of the different solutions, one of the most rigorous personal tests being this one from John Pemberton, which pitted the Libre, Dexcom G5, Medtronic Veo with Minilink and Medtronic 640G with Guardian 2 (R1) against each other. For the benefit of those who haven’t seen the comparative accuracies of these systems, the manufacturers stated levels are below:

Given this accuracy table, you’d expect the recommendations based on accuracy to reflect equally on the Dexcom G4 with 505 (Only available in the US), the G5 (available globally) and the Medtronic Guardian 2 (R2) system. But far fewer people have used this version of Medtronic, so the collective evidence base in much lower. And don’t underestimate the power of bad experiences. This is why you will generally see the Dexcoms (of both types) recommended.

Recommendations also bring into account the cost of the system, and from any perspective, the longer life you can get out of an individual sensor, the better, and the lower the start up cost, the better. On this footing, the Dexcom system nearly always wins. In the UK at least, the start-up cost for G5 is £200 for a transmitter and £55 for the sensor. For the Medtronic it’s around £375 to get started and £55 per sensor with 6 day stated life. And of course, we mustn’t forget the Libre, for which it is now £53 (including postage) per 14 days if you use the LibreLink phone app. There’s a clear difference on start up costs.

Running costs are also notably different. My experience with the various sensor/transmitter set-ups is:

Anecdotally there are people who manage up to 80 days on the Dexcom with a G4 transmitter and 19 with the Medtronic/Guardian 2 configuration. I’ve never seen anything like either of these. If you are paying for your own though, this becomes very important. Even if we use my averages, at, on average, around £50 per sensor, the less frequently you have to buy a new one, the better! If cost is your ultimate driver and you can’t afford the set-up costs, then really, you’re in Libre land. If you can afford the set-up costs, there seems to be cost comparability towards the Dexcom.

From these two sets of data alone, you can see that someone in the US who has only had access to Medtronic using a Veo with Minilink transmitter probably wouldn’t choose the Medtronic solution if all they were interested in was CGM, and this is why the groundswell of opinion tends towards the Dexcom solutions.

Add to this that both Animas Vibe pumps and T:Slim pumps support Dexcom CGM integrated into the pump, and you can further appreciate the benefits of the response from stateside. A few may use the Low Glucose Suspend feature on the more recent US Medtronic pumps, but that seems to be a limited number of users.

But what if I’m in Europe?

Then your options are different. You have access to something that’s only recently become available in the US. The newer transmitter set up from Medtronic. Which is vastly superior, in my experience, to the Minilink. Both in terms of accuracy and following blood glucose levels, but also in range and connectivity. It seems to retain a connection much more effectively than the Minilink. It also has the added benefit of storing the last 10 hours of data, so if you lose connection to the pump, you still have the data (rather like the Libre). This is a benefit that I think is hugely ignored, and one that I found very useful.

And more importantly, for the best part of two years, you’ve also had access to the 640G, which, if you were interested, allowed you to use SmartGuard. And when I say SmartGuard, I mean the full predictive low glucose suspend functionality that I’ve written about extensively before.

At this point you are no longer just choosing CGM. You are choosing something a lot more. It’s not just a pump with an integrated monitor, it’s a pump with a an additional level of “intelligence”. For many people, this is just as important as the CGM on its own. You’re no longer comparing apples with apples.

Of course, the Dexcom G5 also has the “official” sanction for being used to bolus, without a finger prick test, but my experience with the 640G and G2 transmitter was that it was also good enough. Some may not like the official sanctioning, but that’s a choice you can make.

So if you’re sticking within the official framework of devices and systems, I’d suggest it’s always worth asking for a trial of any of the systems available before making a decision to purchase. If I was sticking to the normal set of products, I’d choose the 640G over running a standalone Dexcom, because the benefits are much greater. There is nothing else like Smartguard (until the 670G comes along).

I want to go #wearnotwaiting though….?

This, again, is likely to change your perspective. If you want remote monitoring, then Dexcom is easier to make this work with, but there are tools to do it with Medtronic. If you want to run an artificial pancreas, then the OpenAPS system works with Minilink and Dexcom (and I’ve run both).

Dexcom’s accuracy makes it a better bet, in my book. There are also tools to include the Libre with the NightScout system. You can use all of the above, but most of the work has been done with the Dexcom.

So what would you recommend?

That’s a very difficult question. Whilst I recognise that some people have issues with the Enlite system, I’ve always got on well with it and given the choice of currently available tech, would always go for the Medtronic 640G and Guardian 2 solution if selecting a wholly manufacturer based system, simply because it is so much more than a Pump with Integrated CGM.

For doing the OpenAPS stuff, while I do use Enlite and Minilink, the Dexcom better accuracy really provides better results. If cost is your biggest issue, you’re likely to get the best total cost of ownership using Dexcom G4, but you’ll have to do some of your own soldering to get what you need.

Returning to the original question, what’s the best CGM? It’s the one that does what you need it to. Context is everything.

Whichever you choose to go for, it’s always worth thinking about the experiences of those you have asked for advice. Especially in the context of the CGM world. Globally, all Medtronic experiences are not equal, and neither are all Dexcom G4 experiences. If you’re looking for advice on Medtronic v Dexcom, and you’re getting a pump in Europe, ask the Europeans what they think. You may well get a different answer from those in the US.


  1. I agree with most of the content of the article. I live in the USA and have being on both sides of the fence with dexcom and medtronic. My experience with the the 530G Medtronic back in 2013 was a big disappointment and 2 years later I moved back to Omnipod as a pump and the G5 dexcom. I started with di G4 with the 505software and I had the chance to see the enormous difference both in accuracy and longevity of the G4 vs. the enlite. Today I am running the G5 on xDrip+ and I found the system to be awesome! Perhaps it should had be noticed on the article as another option to evaluate and compare, and non just as a matter of fairness considering the mentioning use of a Nightscout App with the Libre, but also because xDrip brings the experience of the G5 on Android to a new and higher level of satisfaction. it provides much more information and useful data, including that of the use of tool of predictive sugar levels, calculating the interaction of carbs, IOB and values of sensor used with an algorithm. The sugar predictions are shown with a graphic inside the main graphic as well as with notes and alarms. There are an array of other useful functions including the option to have your sugar values being red to you vocally via the speaker of the cell phone leaving you free with both you hands and mind to attend your life. Nightscout and other developers do not make the products that we diabetic use and cherish for our good control, but they definitely have improved most products in a variety of ways and I am thankful to them too.

    • I’d agree it’s pretty good, but I don’t think it’s better than the Dexcom app on iOS. The key point of the article isn’t really the apps, it’s more about whether people realise that asking the question “Which is best” requires an understanding of the experience of the people that are responding and what your use case is.

      My use of Dexcom with OpenAPS is a case in point. I use xDrip+ to achieve that but my use case is rather different than yours. I’d also argue there’s one major limitation of xDrip+ as an app and that’s that it only accounts for IOB from bolusing, where TBRs and basal rates would be useful, but then I guess we are heading into HAPP territory.

    • My problems with Dexcom was loss of signal so most of the time when I left home, I had no signal. Now I am using medtronic’s sensor, I get readings all the time, since it uses Bluetooth.

  2. One thing you don’t talk about is the 20/20, 30/30 and 40/40 figures for the sensors. I know it makes even more complex but to me it’s really critical. The Medtronic sensors, even the latest, seem to have significantly higher 30/30 and 40/40 figures. Even if it’s in the 2% range, it means that sometimes the sensor is going to be off by a *lot*. Now this varies a lot from person to person and lot to lot, but still it’s there. And that’s what would scare me, is that every now & then, even rarely, the sensor is going to be off by a lot. And if you look at it that way rather than the pure MARD, then Dexcom seems a much better choice than Medtronic. Plus the fact that real-life testing of Medtronic sensors shows significantly higher MARD than “official” data (admittedly there is a huge person-to-person variability).

    • Hi Henri, it’s an interesting point, but I’d have to come back to the question I’ve asked in the post. Which Medtronic combination are you talking about here? I’ve not been able to find any data on these figures for the Guardian 2 set-up, although I’d agree the Minilink numbers aren’t great. If you have a source, it would be interesting to see it as I’d be happy to update the article.

      My experience of the Enlite sensors with the Minilink transmitter is that the 20/20 figures are not great. I’ve not checked the personal MARD on them as I’ve no intention of using them long term.

      Using the Guardian 2 updated transmitter, however, was much less of an issue on the 20/20 side for me. It was a whole lot more accurate. Even though they are the same sensor (rather like the Dexcom G4 and G5 differences here in Europe, where we don’t have the 505 Algorithm). Personal anecdotes on the MARD of this system (and my own when I measured it versus fingerprick testing) was that it was performing a lot better than the quoted 9%. I get the feeling that Enlites are more variable between individuals than the Dexcom sensors seem to be though.

      I think that the relevant point for me in the entire article is that we keep finding ourselves in the position where we are about to make what is a significant investment decision, and the mechanisms for trialling these systems are terrible. What we need is the ability to “try before you buy” so that we can make a sensible decision based on what happens with our own body.

  3. You didn’t address the death of the Dex 5 ransmitter @3 months. I wasn’t informed about this when I had decided to go Dex, and returned the whole system after receiving it (although Dex refused to refund the sensors themselves ?). I was transitioning last year to Medicare, and the costs would have been higher with Dex with that feature. I’m currently wearing Enlite, and my transmitter is almost 2 years old. Now the new hitch, Dex is Medicare approved. Hoping MM will follow soon.
    It’s always about profit, and I felt since Dex’s only revenue stream is sensors that was the reason for the kill switch. Now, on FB group pages, I’m reading the Dex sensors are dying sooner and not lasting as long: some have reported that Dex is making cold calls to customers pushing the narrative that Dex sensor is only good for 6-7 days. Oh oh…are they now manipulating the sensors also?
    Anyway, in the last 3 or 4 months, my Enlite have improved AND typically can wear the same one for 10 to 14 days. I’m on day 15 with current one, and still working. I also prefer the lower profile, but that’s a personal decision, as I wear on thigh usually, and much less obtrusive under clothing than the Dex.
    BTW, my son is a T1 also, and uses Dex, so have some experience with them. I confidently micro-bolus off my Enlite numbers. Major boluses, I do a BG test, and what I find is the sensor is lower than BG, if it is climbing, but usually catches up within 1/2 hour or so.
    Not ready for the MM670 yet (target BG is non negotiable, and too high for me). Hoping in later models, customers will be able to choose their targets. But the Guardian looks good.
    Thanks for the article. Enjoyed it, but next time, please consider the transmitter costs also.

    • Thanks Colleen. The main purpose of this article is to ensure that people understand that what they’re discussing and hearing compared when looking at social media is not necessarily what they think they’re hearing due to differences in availability of products worldwide. As a result, the recommendations they may be receiving may need to be contextualized. I leave it up to individuals to go away and determine the individual costs of a system based on their own use.

  4. Oops, couple of typos! Current transmitter is near 2 years old….tried to “edit” but…

  5. Awesome article…. As European (Spanish) I am interested in accuracy of Minilink transmitter with Enlite sensor in order to make Loop (iOS app), I wonder about MARD comparison versus Guardian 2 (which my child is still using with 640g quite satisfactorily) and Dexcom. Officially I guess Dex is better, but what about your experience? TIA

    • Hi Andres, it works okay, but requires a lot more calibration than the Dexcom. I found the Minilink with the 722 to be the most portable option really, and I was comfortable enough bolusing from the readings, but it isn’t as accurate as the Dexcom is. It’s fair to say that you have to get a feel for how the systems work so that you can identify if you feel something isn’t quite right and act on that. Even with the Dexcom G5, we’re still some way from an unattended system.

      • Thanks again for sharing your own and large experience, I am about to start looping with a 754 and Minilink in a 4 years old boy (beginning in open Loop of course)… Also thanks for sharing your Fiasp experience which I find so interesting…So as you write more (and better) than Ken Folett, I think you should get a big salary for that LOL!

  6. I have a Medtronic 640g with Guardian 2 and very mixed results. How do i know if i have a Medtronic Guardian 2 R1 or R2 ?

    • I believe the transmitter part number is slightly different, but if you got your transmitter in the past 6-9 months then it’s highly unlikely to be one of the R1 ones.

    • I’d recommend reading the post again, and deciding based on your own criteria. I personally use the Dexcom G5 because it best fits my use case.

2 Trackbacks / Pingbacks

  1. Nova inzulinska pumpa konačno u Hrvatskoj! - NaInzulinu.com
  2. W poszukiwaniu brakującego ogniwa. Dlaczego Medtronic jest ofiarą swojego sukcesu? | Życie w typie 1

Leave a Reply

Your email address will not be published.