for joining Diabetes in Pregnancy for your last session of the day. It's really been a great program so far. I'm going to share a few highlights from the literature this year and really just a few lessons that I've noted and things that resonated with me in my clinical practice. So here's a picture of the landscape of the west of Ireland which is where I was born and raised and lived until seven years ago when I moved to the United States. So it still has a very special place in my heart and when I consider the landscape of diabetes in pregnancy we've come a long way since insulin was discovered a hundred years ago. At that time diabetes in pregnancy was really limited to people living with type 1 diabetes and I would say that outcomes in general were quite poor. Since then we've seen a big change. There's been a rise of gestational diabetes and even more recently type 2 diabetes has begun to surpass type 1 diabetes in terms of prevalence. And with this we have a whole new set of challenges when it comes to helping women navigate their time before during and after pregnancy. We've also seen the rise of diabetes technologies and furthermore we're also realizing more and more that diabetes in pregnancy is not just the pregnancy itself but we need to take more of a life course approach to this and consider the time period before, during and after pregnancy. And finally clinical trials have become a little bit more common in this area which is welcome. You know traditionally pregnant women were not considered eligible for many clinical trials and people tended to avoid this population due to the complexities but again as you'll see we're seeing more and more trials in this space. So the first thing that I have really begun to understand and apply to my practice is that technology even if it hasn't been developed specifically for pregnancy it can be leveraged to improve outcomes in women with diabetes during this important time. So when we consider traditional glucose monitoring goals most of us will be familiar with that fasting glucose target of 70 to 95 milligrams per deciliter in pregnancy and then striving for a one or two hour postprandial target less than 140 or less than 120 milligrams per deciliter. And certainly if you achieve those goals you will see an A1C in pregnancy of six to six and a half percent or even less. However there are issues with these metrics, these traditional point-of-care testing metrics and the major issue is that complications are still frequent even in cases which would be deemed well controlled by those metrics. There are some very elegant studies looking at functional data analysis on patterns seen in pooled CGM reports and it shows that many patterns associated with large for gestational age, one of the primary complications that we're trying to prevent, these patterns may not be detected on traditional point-of-care glucose monitoring or A1C because they're occurring at time points that are not picked up during the before and after meal times and A1C is not sufficiently elevated to notice any issue. But we don't really even need to go to these studies. Here's just a snapshot of a CGM from a patient that I saw last month in clinic and you can see here that if we were just to do a point-of-care glucose test in the morning we would think that actually this individual was reaching their glucose goal of less than 95 milligrams per deciliter and certainly the A1C is pretty good. But when you look at these strips overnight you can see that she is not reaching her glucose goal for most of the night and this is fueling abnormal fetal growth. So I want to summarize the literature on CGM in pregnancy by saying that in general CGM sensors are safe in pregnancy. They are relatively accurate in their glucose measurements similar to outside of pregnancy and most of the studies would suggest that they are well tolerated in pregnancy and highly acceptable to pregnant women with high user satisfaction. And in fact now the Dexcom G7, Freestyle Libre 2 and 3 are actually approved for use in pregnancy so we have no issues there. And I would also like to say that the benefits of CGM in type 1 and type 2 diabetes that are present outside of pregnancy in the big studies will also be applicable to pregnant women in terms of reducing risk of severe hypoglycemia and so on and so forth. Now the other question is does using CGM improve pregnancy outcomes? And in type 1 diabetes we have of course the concept randomized control trial. And this trial demonstrated very clearly that pregnant users using CGM had better A1C, better time and target range, comparable hypoglycemia, but most importantly improved pregnancy outcomes with arguably quite low numbers needed to treat. Six in the case of large gestational age and neonatal intensive care unit admission and eight to avoid a complication of neonatal hypoglycemia. Women with type 1 diabetes using CGM had a one day shorter hospital stay. So this was really important in the field of type 1 diabetes and certainly has had very significant practice implications. So really across the board all women living with type 1 diabetes should be using a CGM during pregnancy. And this has been recommended by the ADA who suggests that the CGM metrics, things like time and range, time below range, these may be used in addition to looking at the usual pre and post meal targets to achieve optimal control. Furthermore in type 1 diabetes CGM is actually cost-saving compared to using capillary blood glucose testing. Because as you know those strips are not exactly cheap especially in the US and certainly when you're testing 7 to up to 14 times per day in type 1 diabetes using a CGM is more efficient. At this time we don't have any pregnancy outcome data saying that using a CGM in type 2 diabetes or gestational diabetes will improve the pregnancy outcome itself. The other thing about continuous glucose monitoring in pregnancy is there are some controversies about the target range and the time and range. In type 1 diabetes the target range is narrowed for pregnancy and this is fairly well agreed upon. We're aiming for a target range of 63 to 140 milligrams per deciliter and we're tolerating lower blood sugars classifying hypoglycemia as anything below 63 milligrams per deciliter or 3.5 millimoles per liter. However in gestational and type 2 diabetes there is a lack of consensus about the time that you need to spend in that target range for an optimal pregnancy outcome. So people are doing their own thing. Personally I would be striving for 85 to 90 percent time and range among women living with gestational or type 2 diabetes. In general they're a lot less vulnerable to hypoglycemia and they also often carry comorbidities such as excessive BMI which adds to their pregnancy related risks. So we're really striving for you glycemia are as close as possible in these individuals. There have been some exciting advances in CGM in pregnancies without and with gestational diabetes and here is just one paper that I wanted to highlight by Celeste Durnwald and colleagues published this year. And what they wanted to do was assess if CGM metrics could be used to diagnose gestational diabetes or even predict gestational diabetes because as many of us are aware we're currently diagnosing GDM for the most part between 24 and 28 weeks of pregnancy because that's when the hyperglycemia truly becomes manifest on oral glucose tolerance testing. But it's very clear that the fetal growth trajectory has changed already before that time and we have a very short window in which to make an impact and that has really been challenging. So this was an observational non-intervention study in expectant moms without diabetes enrolled prior to 17 weeks. They had a total of over 700 participants and 58 developed gestational diabetes by standard criteria. And what they found was that in early pregnancy those that developed GDM had a higher mean glucose they spent more time greater than 120 or 140 milligrams per deciliter and less time in the range of 63 to 140 milligrams per deciliter when compared to participants without gestational diabetes. And so what they're kind of moving towards is suggesting that CGM derived patterns in early pregnancy may be able to identify individuals who are most likely to develop GDM and provide a wider time frame to make an effective intervention. However we still need to do randomized control trials and there are certainly a number of ongoing trials in this space. I've listed several here and so I'm hoping that in the next few years I'll be able to come back to you and explain if using CGM in GDM and type 2 diabetes actually improves the pregnancy outcome. I mentioned that I typically strive for achieving a time in target range of 65 to 140 milligrams per deciliter of 80 to 90 percent if possible. Some other groups would actually advocate for a more stringent time in target range of 63 to 120 milligrams per deciliter and again I think there's going to be further work comparing the actual target ranges themselves in this space. Moving on to pumps and again this is another area of debate and I think pregnancy is similar to other aspects of diabetes care. The technology is moving so quickly that by the time the trials become available there's another available product on the market which is better and more efficient. Although the concept study was designed primarily to look at continuous glucose monitoring, the cohort of individuals in this were split into people using pumps and people using multiple daily injections. And actually the concept data were interesting because what they found was that pump users actually had a lower time in range and a higher time above range compared to the multiple daily injection users in the concept study. They also had worse outcomes and so hypertensive disorders were higher, risk of neonatal intensive care unit admission was higher, and the risk of neonatal hypoglycemia was higher, borderline higher, among individuals who were using pumps in pregnancy. Now again these data are remember from 2015-2016 when the trial was conducted and at that time hybrid closed loops were not been used in terms of pumps. And I will also say I was part of one of the centers that took part in this trial and I think at the time pumps weren't as freely available and many of the healthcare professionals involved in the care of these women were less confident in making dose adjustments when using pumps. And the women themselves were not as proactive as I see in the current day at making adjustments and trying to strive for those targets. So again these are the best high quality data that we have for pumps and pregnancy which would suggest that if anything pumps may not perform as well as multiple daily injections. But I think this landscape has changed now that we have newer pumps available. And so in the New England Journal last year Helen Murphy and colleagues published on the Cambridge artificial pancreas system. And this is a system, it's an Android app that works with CGM and a pump to have a hybrid closed loop system. And in this trial they looked at women with type 1 diabetes. They were enrolled in early pregnancy 11 weeks gestation on average and I think fairly representative of a real-world population. Their A1c was 7.7% at enrollment and they were randomized either to the hybrid closed loop system or insulin injections and they all had CGM. The primary outcome was percentage time in range from 16 weeks until delivery. This study was not powered to look at things like congenital anomalies or other pregnancy related outcomes. But this was a positive trial. They found that those that were using the hybrid closed loop system had a higher time in range from 16 weeks gestation until delivery. On average 10.5% higher time in range and really not a huge increased risk of adverse events. Particularly no increased risk of diabetic ketoacidosis using this system. And again this is one of the concerns with pumps and pump failures. Will DKA be increased? But this was not the case. I would like to point out that this system is you know available in certain parts of the world. We do not have access to it in the United States although it seems like it's going to be coming soon. This pump hybrid closed loop system has a very customizable glucose target which falls within that pregnancy target. So you can customize it to target a glucose from 80 up to 198 milligrams per deciliter or 4.4 to 11 millimoles per liter. So again a very flexible system. The crystal randomized control trial also looked at individuals with type 1 diabetes and they were enrolled at an average of 10 weeks gestation. Now this used a commercially available pump the Medtronic MiniMed 780G pump and compared the use of this pump to insulin and CGM. These participants were very tightly controlled on admission to the study. So their average A1C was six and a half percent which is different to the previous trial that I showed you. The primary outcome here in this trial again was looking at time and range and what they found was that there was no difference in the hybrid closed loop group compared to the standard insulin therapy group using this pump system. However those using the hybrid closed loop system had increased overnight time and range, decreased time below range and improved treatment satisfaction. So I think this is important and I think speaks to the fact that yes this system targets a fasting glucose of a hundred milligrams per deciliter at its lowest so not perfect for pregnancy but certainly seems to be acceptable among users and we weren't seeing any kind of increased risk compared to a standard approach. So like many of us I'm waiting for fully optimized insulin delivery systems across the board to provide choice to women and in the meantime I think both multiple daily injection and pump are effective but we do need to be conscious that except for the Cambridge APS system the closed loops that are available at least in the US are not approved for use in pregnancy. So I have a few rules of thumb. I discuss with individuals at the onset of pregnancy that the algorithm use is not approved because it's simply not targeting a glucose that's low enough for pregnancy. However, I would kind of caution people that we shouldn't let perfection get in the way of the greater good. And so on many occasions I've found that if the hybrid closed-loop system is turned off, the glycemic control really deteriorates. So we're better having it on and trying to optimize things using the technology rather than turning it off and completely derailing glycemic control. So we should use the lowest glucose target possible. I encourage women to do frequent corrections. This is probably the only time where I recommend using fake carbohydrates, but not more than two hourly. We should use low-glucose suspend features. We can consider super-bolusing, and super-bolusing is a concept whereby you deliver a large bolus for a meal well in advance of the meal in pregnancy because of the delayed onset of insulin action. But then to prevent that slump that you get and that hypoglycemia at three to four hours postprandially, you can switch down the basal rate or even suspend very briefly if needed to try and smoothen out the post-meal control. I would also say that it's important to update the presets very regularly, and this is particularly important in pump systems that have a completely automated basal rate. Because if that individual goes back into manual mode for whatever reason, and they revert to manual mode basal rate settings that were present at the beginning of pregnancy and now we're into, say, trimester three, they could end up hyperglycemic very, very quickly. The other, I suppose, danger point is in the immediate postpartum phase. Again, it's really important that if you're using a pump that has a memory and it's been in an automated basal rate that you put those pumps into manual mode until the memory clears, or else have a hard reset on the pump postpartum. I typically don't start pumps in the middle of pregnancy or at the diagnosis of pregnancy, because again, we have to have education and sometimes that can just cause too much delay. But sometimes I feel like there's no other option. Just to highlight, this is one individual that came to see me at 10 weeks gestation and this was her glucose control. She had a bad history, poor previous pregnancy experiences with pregnancy loss, and really struggled with adherence to therapy. And you can see here this pattern whereby she missed her basal insulin and then took a rapid-acting bolus, which caused hypoglycemia. Then that was followed by hours of hyperglycemia. And so we had a frank discussion and it seemed like she was open to trying something new. So I figured nothing could be worse than what we were currently experiencing. And this is a NomniPod pump. And you can see here that at pregnancy detection, A1C was up on 14% and we improved things by 22 weeks. Still not optimal, but I think better than if we'd continued just the shots throughout pregnancy. So the next point that I wanted to make is that insulin is certainly effective in terms of glucose lowering for gestational diabetes and type 2 diabetes. But in these contexts, metformin should not be forgotten. And I think it has some potential benefit, at least in selected cases. So here's a picture of the pregnancy-safe pharmacotherapy. And you can see that we don't have a huge amount of choice. I've highlighted the insulins here that have randomized control trial data available to support their safety and efficacy in pregnancy. I will say that Lantus or insulin, Glargine, although we don't have a trial in pregnancy, we have a wealth of clinical experience. So if someone's on Lantus at the onset of pregnancy, I wouldn't necessarily stop it. Metformin, kind of an original category B, and Glyburide, category C, are available as options and have been used at various times for treatment of GDM and type 2 diabetes, kind of falling in and out of favor, I would say, over the years. In type 2 diabetes, oral agents are typically not capable of overcoming the insulin resistance of pregnancy. So I think we need to bear this in mind when we're discussing a plan. In general, insulin is going to be required in type 2 diabetes. And sometimes that just makes women think, well, I'm just going to go with insulin anyway from the outset because I'm going to have to do this and may as well get started. Metformin and sulfonylureas can be used. Both cross the placenta. We have one major metformin trial in type 2 diabetes, and we have no major Glyburide trial in type 2 diabetes. So here is the metformin trial. It's the MITEI trial published in the Lancet Diabetes Endocrinology in 2020 by Denise Fegg's group from Canada. And it looked at women with type 2 diabetes on insulin, enrolled at six to 22 weeks gestation, and they added metformin or placebo to insulin. And so what they found was that the metformin-treated group had a slightly better A1C, 5.9 versus 6.1%. So the clinical significance of that can be thrashed out. They did require less insulin. They gained less weight, on average about two kilograms less during the pregnancy, and they had fewer cesarean births. They did find that infants in the metformin group were more likely to be small for gestational age. And this was a concern because is this a direct effect of the metformin? Is this bad? And so it has had a practice implication. We don't really understand what the pathophysiology of this could be, and as I'll discuss, it hasn't been replicated in every trial. But again, if somebody is experiencing small for gestational age on ultrasound or intrauterine growth restriction, I'm generally stopping the metformin. The GI side effects were incidentally similar between both groups. So the guidelines now suggest that insulin is the preferred therapy for type 2 diabetes in pregnancy, but ultimately there will be people who cannot safely administer insulin, those that cannot afford insulin or don't have access to insulin. And in these situations, it is reasonable to discuss metformin and potentially even glyburide. Again, just trying to get some improvement in glycemic control. Now in GDM, we've had some moves in this space over the last few years. Our organizations still give us slightly mixed messages in terms of what's the preferred agent. I would say that insulin tends to win out, but some groups, such as NICE in the UK, recommend metformin as first line in gestational diabetes. Insulin therapy works well to improve hyperglycemia in gestational diabetes, but it certainly doesn't normalize key outcomes to that of the background pregnant population. And so we're always trying to figure out is there some other way to get the outcomes for women with gestational diabetes similar to that of the general pregnancy population. So Janet Rowan's trial, published back in 2008, looked at metformin versus insulin for the treatment of GDM. And these were women who had failed a dietary intervention for GDM, and they were treated with insulin or metformin. The first point to note is that in 44% of those that were treated with metformin, they still needed insulin to be added to achieve glycemic control. They had a similar pregnancy outcome, less gestational weight gain, and women preferred metformin, even the ones that needed insulin later on in the pregnancy. But again, there's always a catch. At two years, higher skin-fold measurements were noted in the offspring follow-up studies of those that were exposed to metformin. No difference in body fat composition at that time or neurodevelopmental outcomes, but at nine years, the metformin offspring are larger by several measures. So again, this has raised concern, and there are other smaller studies and some pathophys studies raising concern that metformin is causing a growth restriction in utero, but then this rebound, excessive growth in childhood. I will say there was a big limitation in this follow-up study, in that a lot of children weren't followed up, so it could be a biased sample, and we certainly need more data. So last year, those of you who were here will remember that Fidel Madon came and spoke to us on the early metformin in gestational diabetes trial. This was a study that I was also part of. 510 participants with gestational diabetes randomized to metformin plus or minus placebo. And so in this study design, we didn't trial and wait for a fail of diet and lifestyle. They were given metformin or placebo at the time of diagnosis. Similar composite pregnancy outcome here, so that was the primary outcome, but metformin was more favorable, again, for outcomes such as maternal weight gain, lower mean birth weight, and a less need for insulin. Now we specifically looked at the small for gestational age infants in a study that was recently published in Diabetologia, and we did not find that metformin was associated with an increased risk of small for gestational age. What we found was that preeclampsia, but not gestational hypertension, was associated with an increased risk in the cohort overall. And metformin exposed small for gestational age infants did not have worse outcomes than SGA infants exposed to placebo. So again, not the same signal that was seen in the type two diabetes trial, and we're hoping to be able to follow up these exposed children to see if there's any effects on their growth as they age. So to summarize GDM pharmacotherapy, I would say that insulin is still considered first line, but metformin is cheap, it's relatively safe, and it may delay or remove the need for insulin therapy, but we need to be cautious with growth restriction. I think we need to consider the use of metformin in non-Caucasian populations, which really, it hasn't been studied to any great extent. We also need to consider the pros and cons of metformin versus insulin in resource-limited populations, where it may be difficult to deliver education on safe insulin administration. And more long-term follow-up studies are needed. Since we've heard a lot about GLP-1-based therapy over the course of today, I wanted to mention the role of these agents in pregnancy. I'm sure many of you have seen women in clinic that are using these medicines to lower their weight, sometimes in an attempt to become pregnant, or improve reproductive health outcomes. This study by Sesta and colleagues is the largest one that's available to date. Population-based data looking at four Nordic countries, Israel and the United States. Over 50,000 women with type 2 diabetes who were pregnant and had a live birth. And they looked at exposure in early pregnancy to insulin and non-insulin agents, including GLP-1 receptor agonists and SGLT2 inhibitors, DPP-4 inhibitors, and sulfonylureas. And what they found was that there was no increased risk of any major congenital malformation or cardiac malformation in the exposed pregnancies. So while this is somewhat reassuring, I will caution by saying these only included live births. And of course, our concern here are animal studies which show that there were higher risks of congenital anomalies among several animal models that were exposed to GLP-1s and SGLT2 inhibitors in early pregnancy. This study will have missed individuals that had a nearly miscarriage or a termination of pregnancy due to a malformation. So again, interpret with caution. This other study published in BMJ Open this year looked at a different way at use of GLP-1 receptor agonists in early pregnancy and reproductive safety. This was looking at teratogenic databases. Again, less numbers in this study, less than 200, but again, no signal of concern in humans. So while the population-based data that we're seeing, convenience samples, real-world data, provide some reassurance to individuals who become unintentionally pregnant while receiving these medications, I still would give caution. The animal studies have noticed an increased risk for malformations using doses comparable to those administered in humans. And really, until further human studies are completed, GLP-1 receptor agonists should be withdrawn the once-weeklies at least two months before pregnancy and SGLT2 inhibitors two weeks prior to pregnancy. And the last thing I wanted to highlight is that maternal health does not end with pregnancy. Many years ago, I looked at pre-gestational diabetes and it was at a time when we'd been doing an intervention to encourage women to engage in prenatal care, pre-pregnancy planning, and had really had great uptake and improvements in glycemic control in our population. And I was quite upset when I looked at 12 months postpartum at this cohort that had been working so well during their pregnancy to find that the average A1c deteriorated at 12 months postpartum to that before pregnancy planning. And even more concerning was that one in five mothers had actually no diabetes care at all in the year following pregnancy, and these are people with type 1 and type 2 diabetes. In the field of gestational diabetes, just one in three in general attend a postpartum oral glucose tolerance test, and there is a 50% GDM recurrence rate and a 60% lifetime risk of type 2 diabetes among people with gestational diabetes. So I think this is a field and an aspect of pregnancy care that we really need to consider and thinking about beyond the birth and how things might be improved. This was a qualitative study published in 2024 that I just wanted to highlight, and they had interviews with 20 high-risk pregnant and postpartum people, and they found that participants really just identified difficulties in accessing primary care postpartum. But those who most successfully engaged had very clear postpartum plans documented in their notes, and I think this is something that we can do if we encounter someone during pregnancy, have a clear postpartum follow-up plan. Good coordination between obstetric and primary care, and I think the models of care where joint obstetric and primary or endocrine care work well, and then access to material resources. So just to summarize, I think this past year we've seen increasing availability of diabetes technologies which can be used to optimize glycemic control in pregnancy, and we've seen a renewed interest in the optimal use of existing glucose-lowering therapies during pregnancy, particularly metformin, although insulin still remains first line in most situations. And I just wanted to give a plug. I was recently lucky to co-edit this current November special edition of Obstetrics and Gynecology. There's a really nice collection of articles there focusing on diabetes and pregnancy, which showcases some of the things that I mentioned in the lecture today. And I'll be happy to take any questions. Thank you.

diabetes in pregnancy

glycemic control

continuous glucose monitoring

metformin

insulin treatment

postpartum care

pregnancy outcomes