"Management of Diabetic Ketoacidosis" by Michael Agus, MD for OPENPediatrics

Just another WordPress site

"Management of Diabetic Ketoacidosis" by Michael Agus, MD for OPENPediatrics

The purpose of this video is to provide general information and education about the care of a critically ill child It is in no way a substitute for the independent decision- making and judgment by a qualified health care professional The information contained in this video should not be used to make a diagnosis or to overrule the advice of a qualified healthcare provider, nor should it be used to provide advice for emergency medical treatment Management of Diabetic Ketoacidosis, by Dr Michael Agus Please note, that in this video we will be following the guidelines used at Boston Children’s Hospital Some of this information may need to be modified based on the equipment, guidelines, and practices in place in your institution Hi and welcome I’ll be speaking today about diabetic ketoacidosis, or DKA My name is Michael Agus, I’m a pediatric intensive care doctor and a pediatric endocrinologist at Boston Children’s Hospital I’m the Director of the Medicine Critical Care Program here Children’s and an assistant professor at Harvard Medical School Hyperglycemia Why does one get hyperglycemic in DKA? Why does the blood sugar go up at all? Well it’s really the same reason you get hyper- anything-emic in any clinical situation It’s some mismatch of increased-ins and decreased-outs So what are the increased-ins of glucose into the blood stream? Well one major one is the one that patients are doing at home What are they drinking at home? In general they’re drinking Pedialyte or juice or some dextrose containing fluid and they don’t realize that they are driving their blood sugar higher and higher The other actor that’s adding glucose to the bloodstream is the liver The liver believes the blood sugar is extremely low because it can’t get any glucose interstitially because of the insulin deficiency and so it turns on gluconeogenesis at full throttle And of course that’s also what produces the ketoacidosis: the lipolysis freeing up those circulating keytones and free fatty acids as well Now where should the glucose be going? Well if insulin were present there would be normal cellular uptake But with a decrease insulin production increased insulin resistance there is a functional severe insulin deficiency and so there is dramatically decreased cellular uptake Now these three factors will give you blood sugars in the mid to high 100s What really pushes you over the edge is when you begin to decrease perfusion to the kidneys We know that the renal threshold ia about 180, and if your blood sugar is anything over 180 and you are adequately hydrated eventually the kidney will dump all that glucose It may get dehydrated in the process and you’ll need to replete those fluids But the glucose should come down In the setting where you begin to hypoperfuse the kidneys, that’s when the glucose really shoots up, and that’s when the glucose will hit 800, 1,000, 1,200 milligram per deciliter range One important caveat to the construct I’ve mentioned here is the middle aged or adolescent patient who happens to be at home rehydrating with water That’s the one patient where the glucose may not be as elevated, the glucose may be only in the 400 range, 500 They may have quite significant diabetic ketoacidosis, however, they are not as dehydrated and the usual patient with DKA These are rare patient but it’s a patient where you really want to think twice when the blood sugar is not that high to ask yourself the question, maybe the kidneys are a little bit better perfused than I’m used to and maybe they don’t need the amount of fluid that I’m about to to give But for most patients they come to the emergency department, they have a blood sugar in say the 1,000 range and they will get their normal saline bolus, usually ten milliliters per kilogram over the first hour And that will almost invariably, significantly drop their blood sugar because we saw on the

prior slide we’re now re-perfusing the kidneys The kidneys are now able to dump glucose And as such, the glucose will begin to plummet After a few hours the trajectory of that glucose concentration will change And now you’re looking at the effect of insulin Insulin will significantly decreases the glucose concentration but it won’t do so at the rate that reperfusion will We know that if we don’t do something about the glucose decrease as a insulin infuses, is it will go all the way down to zero But we maintain a target range about 150 to 250 and in many hands, the appropriate range is really 200 to 300 milligrams per deciliter And we achieve that range by adding dextrose to our IV fluids Now one has to understand that the range here is chosen really to make sure that the glucose doesn’t go severely low A patient is at, in much better shape if they spend the whole night at 250, than if they spend the whole night at 120 If they spend the night at 120 they are at risk of dropping low at any moment, and if not adequately monitored, they can be at significant risk for hypoglycemic seizures As I mentioned at the beginning DKA is a problem of insulin deficiency The cure for DKA is insulin therapy Glucose is really an innocent bystander It is affected by insulin therapy but it’s not part of the problem The problem is insulin deficiency and we need to correct that problem Initial management The normal saline bolus In the emergency department we will generally give 10 milliliters per kilo and that’s been a component of most treatment algorithms throughout most of the literature The real question is do you give that second or third or fourth bolus And it is extremely uncommon to need much volume beyond 10 per kilo and beyond 20 per kilo is extremely rare Now one has to resolve and kind of internally understand that unlike almost every other volume bolus we give in pediatrics, the idea of the volume bolus in DKA is not to restore euvolemia It is to restore end-organ perfusion such that those end-organs don’t suffer significant injury The amount needed to restore end-organ perfusion is very small The amount needed to restore euvolemia is quite significant But if you restored euvolemia in every patient with DKA you would be doing a lot of harm That’s why we restrict fluids to 10 per kilo in general, 20 per kilo rarely But I will say that in the extremely unlikely situation that someone comes in with DKA and shock, they may need 80 per kilo So there is no correct amount of fluid to give a patient with DKA You give the patient what they need But again, it’s not what they need to get to be euvolemic It’s what they need to avoid end-organ dysfunction due to ischemia In terms of insulin therapy, it’s really amazing to remember that the half-life, the serum half-life of IV insulin is between three and five minutes Insulin is very quickly utilized or broken down in serum And so one achieves steady state with a continuous insulin infusion at least a steady state concentration within 3 to 4 half-lives or 15 to 20 minutes So the utility of an insulin bolus was always a little bit in question Recently however though, there was an interesting cohort study in the United Kingdom which demonstrated that any insulin given in the first hour increases the risk of significant cerebral edema quite markedly And so we’ve now shifted away from giving any insulin in the first hour and we certainly don’t give any boluses at this point So based on all this data we proceed by giving our normal saline bolus, as I mentioned usually 10mls per kilo over the first hour and at the end of the first hour we start the insulin fusion without a bolus and without a loading dose We give it at 0.1 units per kilogram per hour as a continuous infusion Fluid and electrolyte therapy Fluid and electrolyte therapy: In terms of choosing what rate of fluid to infuse, it was years ago that we were encouraged to look at a patient and decide what percent dehydration

that particular patient had – this one looks 2.7 percent dehydrated, this one looks 3.4 percent dehydrated We began to realize, based on some nice data, that a handful of expert clinicians would all choose a slightly different percentage when they look at the same patient And so we really don’t have a very good ability to assess total fluid deficit just by looking at a patient, even when we have access to patients’ prior weights There is certainly dehydration that takes place over a period of weeks but there’s also a muscle loss, there’s lipolysis going on, and some of the weight loss may not be attributable to fluid loss and it may not be appropriate to adjust your fluids based on that And so there is now an international consensus statement that guides the therapy that I’ll be telling you about and it specifically refers to the rate of fluid administration and choosing a rate between 1 and a half and twice maintenance In general after getting the appropriate bolus I would start with 1 and a half times maintenance From a dextrose point of view we start with no dextrose in the infusate assuming that the patient is hyperglycemic If it’s the rare patient that I mentioned where blood sugar is 3 to 400 I would put in 5% desxtrose into the solution, but in general we start with none and we ratchet up from 5 to 10 to 12 and a half percent being the highest concentration I would perfuse peripherally Remember that going up on the dextrose should be considered a considered a ratchet Once you go up there is no real reason to go back down As I mentioned, blood sugars in the 300s to 400s are okay in DKA therapy Sure the 150 to 250 is the ideal range but it’s just not worth the expense, the hassle, the risk for errors by changing the fluids more than they absolutely need to be changed So from a dextrose point of view again I’d march up from 0 to D5 to D10 to D12 and a half every time a blood sugar approaches 200 milligrams per deciliter From a sodium point of view, I mentioned to you that there are no data that suggests that sodium administration makes a difference in the risk of significant cerebral edema Despite this there are some anecdotal data which have guided our practice, and the international consensus statement now states that for the first 4 hours we ought to have a sodium concentration in the infusate of 154 millimoles per liter That of course is normal saline And we all agree that for the first four hours we’re going to stick with normal saline At that point however I recommend dropping to half normal saline If we use normal saline the entire course of therapy, every patient is going to end up hyperchloremic with a non- gap acidosis Now non- gap metabolic acidosis is not a dangerous acidosis, but it doesn’t feel quite so good and so those patients may be slower to convert to sub-q insulin, and they may not be in a mood to get all their diabetes education, and they may end up with a longer length of stay because of the hyperchloremia they’ve had in their initial therapy So I try to reduce from normal saline to half normal at that 4-hour mark If the serum sodium is particularly low or particularly high, those are situations where I might continue the normal saline for a little bit longer than the first four- hour mark Now from a potassium point of view, we’ll talk about insulin effects on potassium in a second but, it is clear that patients in DKA are total body potassium depleted And so we think of the baseline potassium concentration infusate for diabetic ketoacidosis to be about 40 milliequivalents per liter The range of acceptable peripheral concentrations is between 20 and 80 When you get up to 80 milliequivalents per liter you may be in a very high range and we’d only do that if the serum potassium is really in the low 2 range and just not responding to the concentrations we are using When one is following potassium one can always look at an electrocardiogram to look for the presence of u waves But in general, hypokalemia, although we certainly try aggressively to avoid it, is not a lethal complication in DKA and one shouldn’t overreact by giving potassium infusions; rather, put it in the IV fluids and the patient will eventually equilibrate over time In our institution we use acetate and phosphate to replace some of the chloride, to avoiding

giving as much chloride as we might otherwise give So instead of giving potassium chloride in the infusate, we give a combination of potassium acetate and potassium phosphate Now potassium acetate can be given in significant concentrations Potassium phosphate, one has to be a little bit concerned about the relative concentration of phosphate and calcium Phosphate and calcium are acutely co-regulated and if one infuses a lot of phosphate the calcium will plummet, and you can develop hypocalcemic tetany as a complication of giving too much phosphate And so we limit the phosphate in the infusate depending on size From a calcium and magnesium point of view we don’t tend to need to replete those at all Metabolic effects of insulin Now let’s review the main effects of insulin We know that insulin pushes intracellularly, glucose, that’s its main function It also has a similar effect on potassium We also know that the patients present total body potassium depleted due to the acidosis In the setting of acidosis cells will try to help to reduce the proton concentration, and that’s basically what acidosis is, right,increased proton concentration The cells will try to help by sucking in a few protons and exchanging them electroneutrally with intracellular potassium As a result that extracellular potassium will begin to rise and when the patient is home and the patient is still well perfused, or the kidneys are still well perfused, the kidneys will dump that potassium And by the time the patient comes into the emergency department they have been dumping that potassium all day long and are total body potassium extremely depleted Now a similar situation happens with phosphate Phosphate can be very low in starving or malnourished people Some believe, have learned about the refeeding syndrome, meaning the risks of refeeding and developing catastrophic hypophosphatemia from anorectic populations The truth is it was first described unfortunately in the Holocaust when allied soldiers freed concentration camps And when they saw these poor, emaciated, tortured prisoners, they said oh you guys haven’t eaten in such a long time, here have a little bit of a candy bar, and the prisoners would eat a little bit of candy bar and many of them would have a cardiac arrest on the spot and die Now why did that happen? It happens because insulin is phosphorylated at the receptor and so insulin uses up a significant amount of free phosphate If your free phosphate concentration is extremely low and you give what amounts to an insulin bolus, meaning its insulin effusion at a significantly high dose which is what we do, you can quickly use up phosphate stores if the circulating phosphate levels are extremely low and all of a sudden you have no phosphate to make ATP from And in that situation you will have instant death of high metabolic tissues like the heart And that will lead obviously to a whole organism death It’s easy to avoid hypophosphatemia, or at least critical hypophosphatemia, by maintaining an infusion of phosphate, as I mentioned in the form of potassium phosphate, somewhere between 15 to 20 milliequvalents per liter if one is infusing the total fluids at between one and a half, and twice maintenance That really is an adequate phosphate delivery in order to avoid critical hypophosphatemia It’s not going to get your patient’s phosphate up to a normal of 4, but it will keep it in the .8 to 1.2 range or potentially above that and that will keep it high enough so that one won’t develop the potentially lethal complications of critical hypophosphatemia Monitoring the patient Now while one is treating a patient with DKA one has to follow frequent lab values We recommend following a blood glucose value hourly That can be a finger stick during the time periods that you’re not drawing an actual blood sample for electrolytes Every two hourly we look at electrolytes, meaning a chem 10, and by that we mean sodium, potassium, chloride, total CO2 or bicarbinate concentration, BUN, creatinine, glucose, calcium,

phosphate and magnesium We usually draw this from a second IV And so we will have an infusing IV with insulin and we will have a second IV where we infuse the IV fluids and we will withdraw from that when appropriate in order to measure glucose concentration and the electrolytes If one is drawing from the line where dextrose was infusing one has to be very careful to have an adequate waste, at least a full mL of blood in order to not contaminate the measurement with a dextrose infusate Another way to measure electrolytes is with a venous blood gas It’s important to remember that although these are valid values they are not directly comparable to serum electrolytes Whole blood values can be a few percentage points off from serum values We also occasionally will measure some form of continuous CO2 monitoring Either end-tidal CO2 or transcutaneous CO2 in order to confirm that the patient’s PCO2 concentration is rising normally It will often start in the very low teen range even as low as a PCO2 of 10 millimeters of mercury But it should rise towards the normal concentration of 40 as the therapy is delivered If that rise is, begins to be flattened or goes in the opposite direction, then something is wrong; usually there is a problem with the insulin delivery system that needs to be relooked at, or occasionally, the insulin concentration was not mixed properly in the pharmacy We conduct a Glasgow coma score every two hours in order to help bedside clinicians identify a change in mental status that might require further evaluation and consideration for therapy Titration of fluid therapy Now there’s two approaches to adjusting fluids in diabetic ketoacidosis The one that we use in our institution involves having a premade set of fluids that we keep in the unit where we predominately care for diabetic ketoacidosis There are four of concentrations of fluid, they’re all normal saline but they have no dextrose, D5, D10 and D12.5, along with a mixture of 25 milliequivalents per liter of potassium acetate and 15 per liter of potassium phosphate By having these four fluids premade and near the patients, as the blood sugar goes down and the provider wants to increase the dextrose delivery, they can order that fluid and that fluid can be hung by the bedside nurse within minutes And using this approach has nearly eradicated the incidence of hypoglycemia during insulin infusion Because this means that the provider will not write the order and leave the unit, the nurse then ends up waiting up to an hour until the pharmacy can make that fluid and increase the dextrose concentration to the patient In the meantime, the patients glucose has been falling steadily And so, we’ve been able to really eradicate hypoglycemia with this method The other approach that many hospitals use is a titration method where they have a high concentration dextrose fluid in one bag and no dextrose in the other bag, and they change the concentration of dextrose by varying the relative infusion rates This makes a lot of sense to me Personally, I have not seen it in action and worry about the training required for bedside nursing and how the orders would need to match in our current situation of computerized order entry In addition to ordering the premade fluid, we also ask the resident or the provider to order the fluid that they really want if it’s different than the store of fluid Maybe they really want d5 for example with 12 of Kphos and 19 of Kacetate That’s fine, but in the meantime they’ll get the premade d5 and then 45 minutes to an hour later they’ll get the fluid that they really wanted from the pharmacy

Now what we do to maintain euglycemia? I mentioned earlier that we increase the dextrose in steps and what I consider it a ratchet So we go up and we don’t go back down There’s just no reason to go back down We start with no dextrose, go to D5, D10, D12.5 Once we hit D12.5 and blood sugar is still in the mid to low 100 range, as in less than 150, then we need to go up on the rate of IV fluid And so we would go up from 1.5 to twice maintenance Now, virtually all patients, the great majority of patients, will finish their course of therapy on D12.5 at twice maintenance If you are on D12.5, if your patient is on D12.5 at twice maintenance and the blood sugar is still going down to the low 100s, then and only then you may need to think about reducing the insulin infusion rate Personally, I don’t advocate decreasing the insulin infusion rate in almost any circumstance because I think that it can introduce error into the system And if a nurse or a provider that is relatively inexperienced believes that insulin is one of the things that they can titrate along with the dextrose fluid rate and all the electrolytes, then the patient may end up not getting the adequate dose of insulin that they need Now what is the adequate dose? I’ve told you that I recommend 0.1 units per kilo per hour Is that definitive, is that what is required? No its not, in fact there’s been some data to suggest that a lower dose may be adequate in many, many patients But I can tell you that 0.1 unit per kilo per hour is adequate in all patients And we’ve never had a case despite having seen hundreds and hundreds of patients with DKA where that dose was inadequate And so I fall back on using an adequate dose for everybody and supporting it with dextrose and the appropriate IV fluid rate in order to maintain euglycemia and resolve the diabetic ketoacidosis 0.1 units per kilo per hour is a massive dose Its 2.4 units per kilo per day The usual diabetic will be on, once they finish their honeymoon period and they’ve settled into diabetes, they will be on about 1 unit per kilo per day So is two and a half times the usual maximum dose for a patient with diabetes But the idea is that when they come in they have a lot of insulin resistance and as that insulin resistance goes away, the insulin effect goes up and up even though the insulin dose is the same And that’s why we need to increase the dextrose supply; we need to increase the fluid supply as the patient moves along That concludes our video on management of diabetic ketoacidosis Thank you Please help us improve the content by providing us with some feedback What did or didn’t you like about this video? Was the content to simple, just right, or too difficult? Was the length too short, just right, or too long? Any additional comments? You can either click the start a new discussion button and type in feedback, or send us an email at [email protected] childrens.harvard.edu Note, feedback is not required to complete this activity in the guided learning pathway