All right thank you so much for coming back after lunch and I'm going to talk about non-diabetes related hypoglycemia. So I'll give you some background and then we'll discuss a few cases and see where that takes us. So just a brief overview glucose is typically very tightly controlled and at around 4.5 millimoles per liter or 80 milligrams per deciliter. The first kind of counter regulatory response occurs and that is suppression of endogenous insulin secretion. And you can see here there in those blue boxes on the graph after that we get counter regulatory hormone release and namely adrenaline glucagon. But then when we get lower down to less than 1.5 millimoles per liter this is when severe neuroglycopenia can occur. 30 milligrams per deciliter is a threshold and typically things like seizures and loss of consciousness can become an issue. So I've listed here some of the autonomic symptoms associated with hypoglycemia and also then symptoms described as neuroglycopenia. In general neuroglycopenic symptoms you know tend to be a bit more specific to hypoglycemia and are highly suggestive. When both autonomic and neuroglycopenic symptoms are present you know that can be a key indicator that hypoglycemia is present. But really it's important to understand that these are non-specific and ultimately before you embark down a route of potentially expensive workup you want to make sure that you've confirmed hypoglycemia. And we do that by trying to identify Whipple's triad. And as a refresher, Whipple's triad is a low plasma glucose, not low glucose on CGM, neuroglycopenia and relief of those symptoms when you correct the hypoglycemia. A few important things that sometimes typically non endocrinologists fall down on. Although a glucose value of 70 milligrams per deciliter or 3.9 millimoles per litre is often used on reference assays, healthy people tend to have blood sugars lower than this and this is just normal physiology. The other point I would make is that continuous glucose monitors are inaccurate at lower glucose levels and should not be really used at this time to confirm a hypoglycemic disorder. I won't dispute their rule their role in helping people that are living with hypoglycemia maybe avoid severe hypoglycemia but they should be not used to confirm the diagnosis. So with hypoglycemia I would say that a good history and examination are very very critical. There are various ways of classifying the causes of hypoglycemia. One way of looking at this is are you in an outpatient setting with a relatively well patient that just has symptoms of hypoglycemia or are you in an inpatient setting where you're dealing with perhaps a patient that's in the intensive care unit on multiple infusions. And I would say in the outpatient setting at Mayo Clinic the most common diagnosis that we end up with our post-bariatric hypoglycemia that has been rising up on the list in recent times. Insulinoma, factitious hypoglycemia which is typically associated with Munchausen syndrome or Munchausen's by proxy. In the inpatient setting we often see drug-induced hypoglycemia and I would say that aside from obvious things like insulin and sulfonylureas other medications like beta blockers and many other meds have been implicated in hypoglycemia. These are usually occurring in the context of a sick patient and someone that's already vulnerable, perhaps malnourished, poor hepatic glycogen stores. So we just need to kind of understand that yes maybe the medication is contributing but there are other factors which may be causing this patient to be at risk. Here is a graphic which just highlights the various mechanisms that can actually lead to hypoglycemia and again this is important when we try and figure out how to treat these individuals. It may be a problem of insulin secretion and again things like insulinoma, sulfonylurea excess, these can lead to excessive insulin secretion. It could be insulin receptor activation so things like non-islet cell tumors that secrete big IGF2 can target the insulin receptor and cause an insulin-like action. Alternatively it could be effects on glucose so things that affect our counter regulatory response and again beta blockers would be an example. They can inhibit hepatic glucose production and that's from inhibition of the adrenergic counter regulation. So again just trying to think you know what might be a cause when someone presents with hypoglycemia. And here are the criteria that we use at Mayo Clinic for endogenous hyperinsulinemic hypoglycemia. So this is really you know if we're trying to figure out is there an insulinoma or something like this present. We generally see an insulin greater than or equal to 3 micro international units per mil, a C-peptide that's at least 0.6 nanograms per milliliter and a high pro-insulin usually greater than equal to 5 picomoles per liter. But this has to be in the context of a low plasma glucose so again sometimes insulin and C-peptide can be elevated appropriately but we want to see these parameters with the low plasma glucose. Sometimes the test can be murky. We love it when they all fit the correct pattern and cross these thresholds but sometimes it's not clear and we try and put several pieces of the jigsaw puzzle together. So what I like to refer to as surrogate markers of insulin secretion are things like beta-hydroxybutyrate concentrations. So if you're fasting and you have been fasting for several hours and yet your ketones are not rising that's usually a sign that there's insulin present in circulation. Likewise if you have been fasting for you know several hours and we bring people in for 72 hour fast, generally you would expect that your hepatic glycogen stores start to get depleted. So if towards the end of a fast you give someone glucagon you really shouldn't expect a very high rise in glucose output after that. So we use the criteria of a glucose increase of more than 25 milligrams per deciliter after 1 milligram of IV glucagon as a surrogate marker of insulin secretion in the preceding hours. We generally want to ensure that there are negative insulin antibodies and a negative screen for oral hypoglycemic agents because again things like sulfonylureas can create a picture that looks like insulinoma biochemically and we need to make sure we have a good assay to rule out these. And so if these symptoms and these biochemistry and lab results cannot be picked up spontaneously then we have to embark on trying to recreate the circumstances that cause hypoglycemia. And we have a very good 72 hour fast protocol and our goal in this fast is to try and confirm hypoglycemia as the cause of the symptoms but also try and determine what the underlying cause is. We usually start people at home the evening before at around 5 p.m. ask them to fast and then bring them to our endocrine testing center early the next morning. After 24 hours of fasting they get transferred to inpatient care if they have not developed hypoglycemia. We allow non-caloric caffeine free beverages so basically water. Non-crucial medications are discontinued and we encourage people to continue normal activity and we do provide a little foot exercise machine in the room to try and keep people entertained. But it's a tough fast and we're doing sampling of plasma glucose, insulin, c-peptide, pro-insulin, beta-hydroxybutyrate every six hours until the glucose falls less than 60 milligrams per deciliter and then we ramp up testing to one to two hourly. The fast is concluded when someone has neuroglycopenic symptoms with a plasma glucose of less than 55 or we also stop the fast once the glucose falls below 45 milligrams per deciliter even if they don't have symptoms and we stop the fast when the person reaches 72 hours if they haven't developed hypoglycemia. At the end of the fast we're doing these hormone concentrations and we're giving one milligram of IV glucagon to try and look at that glucose rise after the fast and as I mentioned a normal glucose rise after 72 hours of fasting is less than 25 milligrams per deciliter. So that's the fast. Sometimes people have more postprandial symptoms and in that case we can do a mixed meal study. Now we kind of differ what we offer for the mixed meal study. In general we're doing this after an overnight fast. If someone is a post-bariatric patient we're generally giving a standardized meal. At Mayo Clinic we have two options. We have a ham egg and cheese sandwich or we have what I call the healthy option which is two slices of whole grain bread with a piece of cheese, small amount of butter and a medium banana. Patients are offered the choice and I'll give you one guess as to which most decide to go with but the idea is that it's a mixed meal and we're trying to see if a mixed meal provokes symptoms in the post-bariatric patients. But if someone does not have a history of post-bariatric hypoglycemia and they are just having spontaneous hypoglycemia after meals we usually ask them to take something that actually is likely to trigger their symptoms and this can kind of vary from person to person. I'll show you an example later on but in this test we do a plasma glucose insulin, c-peptide and pro-insulin at baseline and every 30 minutes for five hours to look at the glycemic patterns. Really important that the OGTT and just a standard oral glucose tolerance test does not have a role in the workup of non-diabetes hypoglycemia. 10% of healthy people will have a plasma glucose of less than 50 milligrams per deciliter during an oral glucose tolerance test and I think this is actually one of the more common referrals that we get. You know women have had for example a postpartum oral glucose tolerance test, have a low two-hour value, feel a little shaky and suddenly end up on this really expensive hypoglycemia workup. Now thinking about some of the causes of hypoglycemia, let's discuss insulinoma just for a few moments. This is the most common functioning neuroendocrine tumor of the pancreas and it's pretty rare. Four per million patient years. What we see is endogenous hyperinsulinemic hypoglycemia with inappropriate insulin concentrations for the prevailing glucose. The 72-hour fast is really critical in diagnosing insulinoma. Within 72 hours of fasting and it's very rare that we say 100% but a large case series from Mayo Clinic, a hundred percent of patients with insulinoma become symptomatic within 72 hours. 20% of insulinomas are misdiagnosed initially as a neurologic or psychiatric disorder. So we have to have a high index of suspicion. 90% of these can be treated and treated very successfully as they are benign solitary tumors of the pancreas. About 6% are associated with MEN1 and some other genetic, rare genetic causes it can result in insulinoma. As I said surgical exclusion is typically curative but there are additional options and just last week we had someone, a 90 year old gentleman who was diagnosed with insulinoma, really not fit for surgery so instead we were going to do chemical ablation at endoscopic ultrasound. Medical therapy, radiation therapy, chemotherapy are really reserved for rarer cases of inoperable or metastatic disease which is kind of a separate topic. Once we've confirmed biochemically the presence or the potential for an insulinoma the localization is where the fun really begins. So CT abdomen and abdominal ultrasound at our center kind of detects about 70% of our patients that can actually then be diagnosed with insulinoma. We use quite a bit of selective arterial calcium stimulation tests and just a quick show of hands how many people have this test available to them? Yeah so scattered and that's what I figured. This is one of these tests that can be really really helpful but you have to have a certain number done at your center every year for it really to be to be useful. But this is really based on the understanding or the observation I guess that when you give IV calcium it stimulates insulin release from insulinoma cells and also from nasidioblastosis but not from normal beta cells. And so what happens during this stimulation test is that our interventional radiologist cannulates the femoral vein and artery and then moves a sampling catheter to the right hepatic vein and uses that to sample insulin periodically throughout the test. And then kind of sequentially again using fluoroscopic guidance a catheter is placed into the splenic the superior mesenteric and gastro duodenal arteries and you inject calcium gluconate into each of these arteries. And what you're trying to see is an uptick in the insulin in a specific artery territory. And so in this example that I have just shown you can see that from baseline at 20 seconds the splenic artery output of insulin goes really high. And in insulinoma usually there is a very high uptick in one specific area. If it's more than 19 times baseline that's very very sensitive for insulinoma. For nasidioblastosis it's usually in one or two territories and you're seeing more like a five or six time increase. So endoscopic ultrasound is another option. It has about 75% sensitivity in our center and then more recently we've got DOTATATE PET which has been reported as quite sensitive for detection of insulinoma. Let's talk about post bariatric postprandial hyperinsulinemic hypoglycemia. So again as I mentioned this is on the rise with more and more people having weight loss surgery. Used to be called late dumping syndrome and initially you know this had been described in about 0.1 to 1% of patients that underwent RUNY. But what we're realizing is that the prevalence is much higher. It's increasing now to about 15% and even up to 30% in some centers. And I would say that some of this is how it's being diagnosed but also how patient selection has been done for these surgeries. The pathogenesis is not fully clear. Generally we think that it's probably a mixed picture. Some of it is from insulin hypersecretion. The symptoms are typically postprandial. I would say that the first line therapy is diet but I do explain to my patients that it's really diet, diet, diet. We want a low GI index diet and again we often use the mixed meal study that I described earlier. I would say that in many of these patients when they're given a standard mixed meal they don't actually become hypoglycemic and we try and use this as a counseling example to say you know if you appropriately mix the nutrients in your meals and avoid things like sugar-sweetened beverages you should not become hypoglycemic. But after that you have to kind of go down a route of trying various medications. Things like a carbose, octreotide, diazoxide and even recently there have been some successes with GLP-1 receptor-based therapy. Sometimes we feed these patients using a tube feed into their remnant stomach or even reversal. Pancreatectomy no longer really has a role in this and there's some controversy about the histological findings in patients that experience this. It had been described that they had like an ascidioblastosis type picture or islet cell hyperplasia. This has been kind of disputed but what we have realized is that the patients that underwent surgery for this procedure more than three-quarters of them relapse after a partial pancreatectomy. And we have to be very careful here you know many of these patients one of the reasons that they had weight loss surgery was to try and treat diabetes and if now we're kind of offering pancreatectomies you know we're kind of trading one disease for another and so we've kind of moved away from that and we're trying other methods now. NIPS, non-insulinoma pancreatogenous hypoglycemia syndrome is not really the same as post-bariatric hypoglycemia. In these patients, and they're very very rare, there is no history of upper GI surgery. We don't understand why they develop endogenous hyperinsulinemic hypoglycemia. They typically get postprandial symptoms. They generally have negative localization studies or the calcium stimulation test might be positive in more than one territory. And the treatment really is diazoxide and oftentimes a partial pancreatectomy. And this is just a picture of the nasidioblastosis that you might see in these cases. Insulin autoimmune hypoglycemia syndrome is clinically and biochemically very similar to insulinoma in terms of its presentation but here we're seeing very very high concentrations of plasma insulin and insulin antibodies. It's more common in people of Asian ethnicity. It's associated with autoimmune disease and HLA-DR4 positivity. What happens is that antibodies bind to insulin and in fact pro-insulin and you get hyperglycemia followed by hypoglycemia when these antibodies spontaneously dissociate or potentially the insulin secretion actually just exceeds the antibody binding capacity. These individuals again rare and so we're reliant on kind of cases on which to base our management. Steroids tend to be quite effective. Immunosuppressants have been used in various forms but it tends to be actually self-limiting this disease and tends to settle down after a period of time. Non-islet cell tumors are rare. They're associated with recurrent hypoglycemia and they can occur in benign or malignant solid tumors and the origin of these tumors can can vary but what we see is big IGF2 binding to the insulin and IGF receptors and so you see hypoinsulinemia in these individuals with low C-peptides in the context of hypoglycemia. Usually we see growth hormone and IGF1 low. So of course the treatment here is to try and identify and treat and excise the tumor if possible. Sometimes you know this occurs in the context of very in-stage metastatic disease and we're trying things like diazoxides, somatostatin analogues, steroids. Imatinib has been described as used in these cases. Hypoglycemia and concurrent illness occurs in the context of decreased glycogen stores, impaired gluconeogenesis and increased peripheral glucose utilization. I would say this is a poor prognostic marker in general if someone's in hospital and they're developing hypoglycemia. Adrenal insufficiency is something that you don't want to miss. I would say that in adults it's actually rare for them to present with hypoglycemia as the primary kind of presenting factor or feature of their adrenal insufficiency, more common in children. It's also present in secondary adrenal insufficiency and you know again these cases you see immediate correction with glucose or in the glucose once people get glucocorticoid replacement therapy and so you know that can be a very satisfying thing to diagnose and treat. So for the next few minutes I'm going to go through a few cases and these are real cases and so again maybe not the prettiest picture in some of the cases but this is a 25 year old woman. She has a two-year history of spells initially once weekly, lasting 15 to 30 minutes. There were no specific triggers that she could identify and she described palpitations, lightheadedness, sweating and this phenomenon where she felt completely dissociated from the environment around her. Initially they were occurring once weekly as I mentioned but they ramped up with time and started occurring almost every day. She'd been unable to work as a nurse for the past two months so this was obviously significantly impairing her life. Initially she said that the symptoms had improved with food intake but this was really not consistent over time. She had been adhering to a very low carbohydrate diet. Her BMI was 22 but she had been just excluding carbs from her diet. She wasn't really clear on why she was doing this but she said that she thought it would be healthy. She saw a psychiatrist and the psychiatrist had diagnosed a possible anxiety disorder but said there was atypical features. They wondered was she having some form of absence seizures and so they tried her with gabapentin. No other past medical or surgical history, she's not on any medications and she had a normal physical exam and she was actually sent to Mayo Clinic to evaluate for a seizure disorder. So she was admitted to the epilepsy monitoring unit. Baseline labs were normal including a morning cortisol that's part of their protocol and she started a sleep-deprived EEG. After about 16 hours 2 a.m. in the morning some blood work was done and as part of the blood work a plasma glucose was completed and it was 38 milligrams per deciliter and during that time they captured EEG and it just showed these non-specific encephalopathic changes. So I actually happen to be covering endocrine inpatient service so the next day I got a phone call about this patient. So let's see a show of hands. At this point do you think that this 25 year old nurse has an insulinoma? Hands. Cerebral petition insulin use? Okay we're seeing more hands. Hypoglycemia associated with a systemic disease? Anybody? What about D? Something's wrong but you're not sure what. Okay good. All right so the patient was already an inpatient and neurology were really not happy with this situation. They wanted her on a different service so we convinced inpatient medicine to take her over and we said we would just start fasting because we just didn't have anything more than a glucose and as it turned out we couldn't add on the samples that we wanted to add on. So we started a 24-hour fast. This was great. 24 hours in her glucose was 83. 30 hours in it was 88. 36 hours in she had a plasma glucose of 71. Ketones were not rising at 36 hours which is kind of curious and then at 40 hours the nurse walks into the room and the patient is barely responsive. So they did a point-of-care glucose and it was 16 milligrams per deciliter. Immediately a blood draw was done. Her plasma glucose was 28 or 1.6. She had an insulin that was pretty high 56.3. C-peptide was elevated. Ketones were completely suppressed and look at her pro-insulin. It was through the roof. She had her one milligram of glucagon at the end of the fast and she rose which is inappropriate for someone that's been fasting this long. She had a sulfonylurea screen that was negative and she had insulin antibodies that were negative. So I won't ask you why you were suspecting that this was cerepetitious insulin use but you can see here that the biochemistry is more in favor of an insulinoma. And so we started a workup and then this was interesting. She had a CT done. The report and our radiologists are good. My husband's a radiologist so I think I've got an extra set of eyes on these scans. Nothing that they really wanted to call a tumor. They had noticed a small focus of ill-defined hyper enhancement in the pancreatic head but was really of uncertain significance and I've pointed out here but you know when you see this in a series of images it's very hard to send someone to surgery on the basis of this and our surgeons like to have a target. So we did an endoscopic ultrasound and again it was just really what we would say non-diagnostic. They were talking about this very vague indeterminate region within the pancreas head but they weren't really saying it was in the exact same location as the CT. So again back to kind of more investigations. So we did our selective arterial calcium stimulation test and remember what we're looking for is a very high increase in the area of one of these blood vessels. And again you can see here that in the gastrojudenal artery the baseline was higher but the rise was not in keeping with insulinoma. And insulin is pulsatile remember so sometimes you do see like a difference in the basal insulin you know when we're doing this testing. So again this is a non-diagnostic study. You can imagine the cost now is accumulating as well as we continue on this workup. And at this point you know I had spoken to the surgeon and said you know are we going to do an exploration or what's next. So we did a pet dotatate scan and this actually showed a tiny focus of FDG uptake in the medial pancreatic head which they said could represent a neuroendocrine tumor. And this was you know in the same location as where the CT had the abnormality. We had another look at the endoscopic ultrasound and this girl went for surgery. Interoperative ultrasound was used. They found a one centimeter hypoechoic nodule in the head of the pancreas. Underwent successful enucleation. Pathology showed a well differentiated neuroendocrine tumor. Six months later she's back at work absolutely no symptoms and just doing really well and complete change in her life and it was just a very satisfying case. But I have to say the workup sometimes makes me makes me nervous especially in these younger patients. Case number two, a twenty-five or 26 year old male, excuse me, BMI of 25 kilograms per meter squared. He had a seizure disorder that was diagnosed six years prior. Six years. This manifests as him taking the fetal position. He has two to three hours of blurred vision after the event and he's treated with phenytoin and gabapentin. He then had an episode of a tonic-clonic seizure despite having therapeutic levels of these medications. Fifteen days after his first tonic-clonic seizure he woke up with palpitations around 2 or 3 a.m. He didn't feel right and his glucose was checked by parents. Sister has type 1 diabetes and it was found that his glucose was 38 milligrams per deciliter. Orange juice was taken and when paramedics came to the house his glucose was actually in the 20s on a capillary blood glucose machine. He was given IV dextrose and he was dismissed home but then he had several further episodes over the next six months some of which were associated with seizures. There was no real pattern in relation to food but his glucose was low on the point of care and improved with juice. He had a 72 hour fast in an outside facility which was called normal but they didn't do the end of fast procedures with the with the glucagon. And he was referred to Mayo Clinic for further exam examination. Physical exam was unremarkable. Okay show of hands do you think he has insulinoma? Cerepetitious insulin use? Hypoglycemia associated with a systemic disease? Okay well a lot of you must want to go for D then. Something wrong but not sure what? Okay all right. Baseline labs were normal including a morning cortisol. Insulin antibodies were negative. Fasting glucose was normal not negative. It was a normal fasting glucose. So we did a 72 hour fast and remember you know this is putting this individual through a second 72 hour fast but again you know he was here for a second opinion and the other fast hadn't had the end of fast procedure so we said we'd go for it. All right so 22 hours his plasma glucose is 88. Ketones are 0.6 at 28 hours. We're seeing maybe a slight rise in beta hydroxybutyrate. Expect this is normal. Plasma glucose was holding its own. 31 hours in comes the nurse and the plasma glucose was just low. I can't get a number. And we get the venous labs and it's 28 milligrams per deciliter 1.5 millimoles per liter and here we have undetectable insulin. Less than 0.5. C-peptide is low. Beta hydroxybutyrate is low. Pro-insulin is low. The end of fast glucagon is flat. He doesn't have a rise and his sulfonylurea screen is negative. So do we think this person has insulinoma? Do we think this person is injecting insulin? I've got one yes in the room. Okay so I want to talk to you about this and this is where the workup can become quite tricky when we start to consider assays and assay limitations. So just to hone in here the plasma glucose was 28 milligrams per deciliter and insulin was undetectable. You always need to know your assays. This insulin was done on a rash platform and if we look here we have several options that are commercially available and they detect analog insulin with varying success. They will detect human insulin or NPH or regular insulin in the same way but depending on what analog is in circulation you may actually not pick up the analog and of course if someone has injected an analog of insulin the human insulin should be suppressed. So glulisine in particular is not picked up well by any of the assays and so at Mayo Clinic we developed an insulin analog assay using mass spec and we use this assay in cases where we get unclear testing and in this man's case when we ran for the kind of five common analogs we found that glulisine levels were through the roof and lo and behold his sister did use glulisine for treatment of her type 1 diabetes. So the patient strongly denied using exogenous insulin and we tried to have an open discussion and eventually it was kind of suggested that he would meet with psychology and our psychology colleagues were able to you know kind of discuss with him the issues and he underwent cognitive behavioral therapy to good effect. So just very quickly the last case is a 41 year old female who had a Roux-en-Y gastric bypass eight months prior. She presents for evaluation of hypoglycemia, typical spells, she's been having lots of outside labs, glucose is in the 50s and in the 60s, we've tried to modify a diet it didn't work and I just really put this up to kind of show that you know these individuals are running low average blood sugars on CGM, they're often dropping down into the hypoglycemic range. She had postprandial symptoms while walking in the subway system at Mayo Clinic, had a seizure just after labs were drawn, we weren't really picking up very low glucoses and but we did get one of 52 milligrams per deciliter. This is her mixed meal, really not seeing a big drop when you eat a standard meal. Her sensor was showing hypoglycemia during the mixed meal but she was not getting biochemical hypoglycemia. We tried octreotide, no improvement. We sent her to the surgeons they said she needed to increase her BMI before they'd consider revision surgery and so she was treated with nasojejunal feeding and decided not to proceed with reversal. So these are tricky cases and we have to manage them on a case-by-case basis. So in summary I think you need to establish Whipple's triad before making a diagnosis of hypoglycemia. You want to treat acute hypoglycemia and try and figure out the underlying cause for a definitive treatment. Just want to thank and acknowledge Dr. John Servus who did a lot of these studies at Mayo Clinic and unfortunately is now deceased. Dr. Adrian Vela and Dr. Pankaj Shah, my mentors in the hypoglycemia practice. Dr. Jim Andrews who has done a lot of our IR studies and Dr. Travis McKenzie who does the surgeries for us at Mayo. Thank you.

hypoglycemia

Whipple's triad

insulinoma

diagnostic protocols

autonomic symptoms

neuroglycopenic responses

non-diabetic